{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"**BÁO CÁO BÀI TẬP LỚN HỌC MÁY CUỐI KỲ**\n\n**Mã Lớp**: INT3405_1\n\n**Họ và tên**: Phạm Văn Hệ\n\n**Mã số sinh viên**:18020468","metadata":{}},{"cell_type":"markdown","source":"# **I.Mô tả bài toán**\n**1. Đặt vấn đề**\n\nQuora là một nền tảng cho phép mọi người học hỏi lẫn nhau. Trên Quora, mọi người có thể đặt câu hỏi và kết nối với những người khác, những người đóng góp thông tin chi tiết độc đáo và câu trả lời chất lượng.Vì là một nền tảng mở ai cũng có thể đọc và hỏi đáp một cách dễ dàng nên sẽ có những người đưa ra những định kiến, những câu hỏi mang tính độc hại chia rẽ. \n\nThách thức ở đây là làm sao loại bỏ những câu hỏi thiếu chân thành - những câu hỏi được đặt ra dựa trên những tiền đề sai lầm hoặc có ý định đưa ra một tuyên bố hơn là tìm kiếm những câu trả lời hữu ích, qua đó cải thiện nền tảng để nó thành 1 nơi mà mọi người dễ dàng trao đổi hơn không bị ảnh hưởng bởi những toxic question như vậy nữa\n\ninput: câu hỏi dạng text\n\noutput: 1/0(không chân thành/ chân thành)","metadata":{}},{"cell_type":"markdown","source":"# **Phân tích dữ liệu**","metadata":{}},{"cell_type":"markdown","source":"đầu tiên import các thư viện","metadata":{}},{"cell_type":"code","source":"# Import các thư viện cần thiết\nimport os\nimport time\nimport numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\nfrom tqdm import tqdm\nimport math\nfrom sklearn.model_selection import train_test_split\nfrom sklearn import metrics\n\nfrom keras.preprocessing.text import Tokenizer\nfrom keras.preprocessing.sequence import pad_sequences\nfrom keras.layers import Dense, Input, LSTM, Embedding, Dropout, Activation, Conv1D, LSTM, GRU\nfrom keras.layers import Bidirectional, GlobalMaxPool1D\nfrom keras.models import Model\nfrom keras import initializers, regularizers, constraints, optimizers, layers\nimport matplotlib.pyplot as plt\nimport seaborn as sns\nfrom nltk.stem import PorterStemmer\nfrom nltk.tokenize import word_tokenize\nfrom nltk.corpus import stopwords","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","execution":{"iopub.status.busy":"2022-01-07T00:23:56.085867Z","iopub.execute_input":"2022-01-07T00:23:56.086217Z","iopub.status.idle":"2022-01-07T00:24:02.102437Z","shell.execute_reply.started":"2022-01-07T00:23:56.086126Z","shell.execute_reply":"2022-01-07T00:24:02.101687Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# chuẩn bị dữ liệu","metadata":{}},{"cell_type":"code","source":"train_df = pd.read_csv(\"../input/quora-insincere-questions-classification/train.csv\")\ntest_df = pd.read_csv(\"../input/quora-insincere-questions-classification/test.csv\")\ntrain_df.info()\nprint(\"Train shape : \",train_df.shape)\nprint(\"Test shape : \",test_df.shape)\n# train_df.head()\n","metadata":{"execution":{"iopub.status.busy":"2022-01-07T00:24:02.104695Z","iopub.execute_input":"2022-01-07T00:24:02.105082Z","iopub.status.idle":"2022-01-07T00:24:08.383182Z","shell.execute_reply.started":"2022-01-07T00:24:02.105045Z","shell.execute_reply":"2022-01-07T00:24:08.382358Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"****","metadata":{}},{"cell_type":"markdown","source":"# Đánh giá dữ liệu","metadata":{}},{"cell_type":"code","source":"\ntarget = train_df['target']\ntarget_1 = 0\nfor target_value in target:\n    if target_value == 1:\n        target_1 += 1\nprint(\"Số câu hỏi trong tệp train:\", len(target))\nprint(\"Số câu hỏi được gán nhãn là 1:\", target_1)\nmyLabels = [\"insincere question\", \"sincere question\"]\nmyCounts = [target_1, len(target) - target_1]\nplt.pie(myCounts, labels = myLabels, autopct='%1.1f%%', shadow=True, startangle=90)\nplt.axis('equal')\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:35:12.072332Z","iopub.execute_input":"2022-01-06T17:35:12.073099Z","iopub.status.idle":"2022-01-06T17:35:12.432884Z","shell.execute_reply.started":"2022-01-06T17:35:12.073056Z","shell.execute_reply":"2022-01-06T17:35:12.432148Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"**Đánh giá nhìn nhận qua:**\n\ndữ liệu gồm 3 cột thông tin:\n\nqid: ID câu hỏi\n\nquestion_text: nội dung câu hỏi-> cần được phân loại\n\ntarget: kết quả của các câu hỏi:0/1(chân thành/ không chân thành)\n\nDữ liệu gồm 2 file tập train và test trong đó tập train gồm 1306122 câu hỏi, tập test gồm 375806 câu hỏi\n\nLớp dữ liệu có sự chênh lệch rất lớn giữa câu chân thành và không chân thành. Tỉ lệ 2 nhóm 1:15 nên nếu sử dụng phương pháp tính accuracy thì chỉ cần đoán tất cả là 0 thì tỉ lệ accuracy đã lên tới 94%.Mục tiêu của bài toán là tìm ra được các câu hỏi không chân thành qua đó loại borb chúng, cho nên để giải quyết vấn đề này ta có 2 cách khá đơn giản: tăng số lượng không câu chân thành lên hoặc giảm câu chân thành xuống. Cả 2 phương pháp này 1 cái làm mất dữ liệu khiến cho tính bao quát của mô hình nhỏ đi, còn 1 bên thì lại khiến ta đau đầu tìm thêm dữ liệu=> không khả quan dẫn tới cần 1 phương pháp đánh giá khác và bài toán này đã sử dụng F1_score làm chỉ số đánh giá mô hình","metadata":{}},{"cell_type":"markdown","source":"# F1_score\n\n![pr.png](attachment:dbd05238-5942-465b-aa18-6161a653d3eb.png)\nPrecision được định nghĩa là tỉ lệ số điểm Positive mô hình dự đoán đúng trên tổng số điểm mô hình dự đoán là Positive.Precision cầng cao tức là số điểm dự đoán positve càng đúng không bị nhầm lẫn\n\nRecall được định nghĩa là tỉ lệ số điểm Positive mô hình dự đoán đúng trên tổng số điểm thật sự là Positive (hay tổng số điểm được gán nhãn là Positive ban đầu). Recall càng cao tức là số lượng điểm positive bị bỏ lở càng ít.\nTuy nhiên, chỉ có Precision hay chỉ có Recall thì không đánh giá được chất lượng mô hình.\n\nChỉ dùng Precision, mô hình chỉ đưa ra dự đoán cho một điểm mà nó chắc chắn nhất. Khi đó Precision = 1, tuy nhiên ta không thể nói là mô hình này tốt.\nChỉ dùng Recall, nếu mô hình dự đoán tất cả các điểm đều là positive. Khi đó Recall = 1, tuy nhiên ta cũng không thể nói đây là mô hình tốt.\nF1 score là trung bình điều hòa giữa precision(dộ chính xác) và recall(độ bao phủ) \n$$Precision = \\frac{TP}{TP+FP}$$\n$$Recall = \\frac{TP}{TP+FN}$$\n$$F1 = \\frac{2}{Recall^{-1} + Precision^{-1}}$$","metadata":{},"attachments":{"dbd05238-5942-465b-aa18-6161a653d3eb.png":{"image/png":"iVBORw0KGgoAAAANSUhEUgAAAo8AAAF/CAYAAAAo6iP/AACvhElEQVR42uzdd3wT5R/A8U+apnsvym6BMsrelCFQQBEE2VP2EAVFVFTEgQwZIv5AQWTIkikbBREF2QIqe08pq4tCge4mvz/uqYRSaFpW037fvPIqSZ5c7p67y33vmSCEEEIIIYQQQgghhBBCCCGEEEIIIYQQQgghhBBCCCGEyPNsgWlAafW8EjBJvZ4TFQJ6AK6ZpKsMdJDdK4QQQuQIBsARsJGssH72wE2ggXreEghTr+dEzwM3gIBM0n0MnJTdm6vYAaXMbmwcgZLyQySEEDkurqgCDAa+A7aq6/FVIAKoI1n0dA1DK1HLy8GjAfBIFzCMA4qlS+cAuMshk6sEASlAPvU8BLgFuOSy7WwAXAYCZZcLIayEjQoKp6nfr1TAhA6Tk4+TyTfY1+Re1N0EmMzijRzBNg/snFYqet+fhw/QZLSSR3NdgNXAObPXEtRDPBsFgdvqxkQ8mDvgDFwxe+2i+gGWvBNC5HR6tEKn94HqgI1PGR+KP1+cgAYB5K+SHydfJ2wdbNk9cTeb3tuUI6PenM4BeA/YDhwBpqKVopmrDCwFjgHb0Ip9i6IV/QYAr6r/91LpqwNfpAuePwHqmj1voYKrY8DPQFUL17cqMDHdsnXqIHnYnUMd4DO1Dt8DG4HX0u0jG6AfsBP4R6V3SRcQbgWOAxuA1ur1YsBktJLRciovvFS+fmeWrrF6DaA5MDSD42UCUF49rwKsBQ6r5Uup5aP5Fugo2ZCpTiqvzJ0DxgDXJXuEEDlYJWALsEJvp69ZqVclmz67+vDaoddo+r+mlG5VGvci7hgcDeh0uhy7EdYQPHqjdfh4D2iP1l5rRrrAcZMK8noC49VrccDfaMXAZ4G/gPPqM8WB7mYBnk4tu5TZcusCc4GXgR0qGPO2YH3PAr1VIJYmABiuSkcepIwKMF9Vwepp4Gu1TWnGAwOAj4D+KhBcjdaurRIwBRil1nmaeh20KsveaNXXN4GDgBE4ofLlskpXAa2kFlWq8wlaaViammo551UAuQlYD3RF64zztfwu4Ae0BToD/g+443xO7ddQtY9s0apbXQAfFeynfdZRHT/m52r+dDdQ9kAjtE5RWWmi4QIUUce/OV+1Hg/ioc4VnToGGwBOGaSrAHRTf9MLABoCbmobq3JvlXMl9dmm6rhNU1Ctn7PKp2JqPezN8sn5Advloz5rvq86AS+k+w4hhHjc9CoO2KPT6+pV6llJN+jkIF7+/mUKhRTCxlaaoT9poSowdFPP12VQCmEeHB8H2mRQchGOVqqZFjweBvo8YDm2aA1WO5ldrB/W5nGiCqrSfAasyWS7+qrA0/yCNwutut1GXazvoJWopnEAzqjg7UW0hrX5Mlh2+nZujmp7aqVL9zawy+xA/xMYYvb+TGC6+v9SFczaqIejykOfPHxs1laB+FL1iFR5bx687AZOAT+qm5vf1c3RCXVch6t9+r36TCMgNt2NyxbgA7P9dBz4FZgNXAI+NUv7sDaPJdT/K5ql16GV8Pd4yHZ2Bf5V2/i7utE4bRbQ2gBfqjQL1N93zT7/rjrW16j8+kedx2nn32TgkMqDvWi1CY7qvbVAFFr1/hn10KPVJtxRNzFB6v1S6db7L3XThTpfrqn9sFNth4P8vAohngAv9Rtt8g32NfXa3sv0qYX/mkxoIm0eH4FOlRIEq4BHry5+qUAN7i2JfJxc0EoEC6t18LDwc98CR9FKOC+qi+2bFnwuWR0kaVagDZ3jpkqrLqkgNU2CCvYaAm8BF9RFfJ86UKcA8dnc9lRgngpqp6jSnFZAM/V+HXXBfsHsM4VUiU9UHv2BOK6Ol1j1fAowGq0U2qSeRwL1gSSVppjar6XRSrgXoDUlsJQRrXTuX/W8Plop+Szulig/yBnuNo8YoF6rqwLNZZl8tghas4ZlKv1ZtZ3L0Zp8tEUrqb6m3t+tjieDCm7LqO0eoG7KWpgtexJa05O0c/CkOg/mqRu1D9Tx1+Ih2/UnWjOOtEC6qgoql6rzeI767tXq92SBWo+5cp0TQjxGhYF16KhYoWsFmn3TDHt3e6vfKGsIHougVcH6qxIKmwwCS9Nj/k4DWmlhK+CAujBnpVrrLPATWhvLTWr9stPi1ag+a1IXOAP3V8XZq5Kl22ilsk1VCdggoCxa9Xx2/aiCn0rqYh+m9kFacDlclQqZi8zDPxIx6Z7vR2tPqlfBd1sVYCWZpTn3iN9pMgscQSvBBK0K+LIFn/8aWIU2TFMkWunfDxbcdNxCKy1M2+eXuFuy3UXdPDU1S39F3ehdV5+9ZLa+gerYTlavmW9PnAoeg7OYJ98DI9DaQSahVYGvRKsx6KDOI0/uNguJUIG3BI9CiMfFH/hNp9eVrDesHg1GNECn1+WKDbOG4HGEKr1oqYKpEHUhAK2a6i+0Nn4/mwWRtiq4edB2pg82bbi3WrAJ8ApaW60b6kLzShbXewpaNa87WnVisoVBa1owrEOrbj+uLrY70dp7VeBuz3FPoJ4qibFBK4lcpR4bLCg9yiwgjlL52lkFonPN8nULWmno4nT5aMrDPxQGdSy2Vnebhc1udvzU//99At9bSQVB5dUxYY/l7Zl3oFWjd0QrfWuJZeOJ3Ul3jhnN/l9ABX3lzF77UwWZx9W6vahuPNqov+bnhx9a29q6aFX9wWhV/FnxE1rVeV1gj9q+tAHw86v1L2uWPkXlhRBCPA7uwFqdTlcydHQodd+ve3/RjwSPT1QUWtVoQ3Wxei/d+8OBX1S6X9FKKntyt13jNXWBuo3W2eZ/qpTBHa1a7RJaCaF5Q/rr6iLcBq39VzN1wcmK7WjVly3QSgItUViVlKxRAUg3ddEzorUnnIpWLfiRCijfVSU3y9AGAu+kSlduqTw49IDvSVDb2EltZz4V6GZkDnerG3uavT5SBZCj0dqhFVKlVn3R2l7mRV+qgGmE2i8vq/+jjsUUtHaGFx9DkJqmsdpH76nv16N1HrNUqjquXlPHxH7ull5m1xl1Pg7N4D1XFUSOQSuNPYrWSQyzG6I/VSD7rjqW1mRjHWLRmn10U+f6dbSqc9BqBlLUTVeKXOOEEI+ZTv2uVq81pBZ13quTqwJHawkeJ6rAZDRa9dVcFfDFqff3qwvou2gdOMLUTktrdzcerd1TX7T2TqCV4s1GG3bmPFo13QS09oKgNdIfg9a4PgKteu5Ls4tqqvrMNfX8ogrgzEtiTOoCGMK949E9zHkVJI5RzzurEpQ0Q9X7vdAa929EK+FMQGvnWAV4A62N5BHurZL7wexCaULrlT0IrcR1nnr9KHc7JpiXTK1AqwINN3v9nAro30er+ryigpiIPPyDUVfl5Ta00rWgdMHMInUs7ker4nZSN0br1L5JQCu1Q/29AiRyt8kC6lwINjsuaqhg7Ue1jNrc7WVvqZUqyO2L1iHqUUuPZ6ubnCVoJYY6dWxeV9uSpG6q7FS+3Ex3A5Ufrd3wNRX4Fcjg5sdf5bEvD25jO1fd2Hip/ZJ2/P+ubsjeQmtfaVQ3UKW4vxmGEEJkVQ+gS/Hni9N4fGN0Nrpct4E62cdP1BJ1Mf7DgrR9VQBcWrLNag0GPkdrF2qrAqWCaD3kU1RQPw+txDtMBUW70TpUJaO1O3wfrRo5Gq35hLe6Ybiobi7s0HpHz0KbJagMWqepyyqI0qmA8gUVCAWhVRUXVMF/CFoJfX600vg0H6ONIVqKzNs7dlU3dUVVIGijblYmqIBNp25wRqKV8rmrNG3VzcU5dfOTqgLVcHXD97XKs61o7SCPqQC7gLo5SyvJrKnOqZNmAbiPes0freQd9Z1/qm0qy912lqjnC9W63lRB60h1AySEENmVHzjg7OfsN+DgAFz8H20yr11f7EobJLyhhbHEU2Er+/mxs1MXMj+0Kkopycg7vlYnt6sKAGPROpGklXjFqgCqpAqIzqogMq294Dj1+evcLQWPRiu1K6MCxHNq+Ynq/eNoVeVBKjj6F603cdqMQufUd6WVzv2F1sP7Trp1D8eyjjKglURv5G47RSNa29u0ZaZ1WFmj1u2GWs8kFSRuQRtDNFX9BtUEvlJB93a0moTqKqg7qc4p81L9PeozTmil5WkdcALSBcRGtSw7lY/mjqplVFDvH+P+Dk9CCJFVnwJ+jcc1fuTAMSeT4PHxc0DrrKJHq/Y2Wvi5y9wdY1FYJyPaAOzmEjJIc4KM2xUmq+ApvWvcbSJBBkHOZe7tWW3e4z2Ve5sSJHN/j3g9Wq/8HhZuZ0bTWEZnkC4arRTRXEm0kk/zdpln0dpsprU7jkv3uYwC2kMZ5F1GPf1jH7IdiWjNPYQQ4nEoAXQvVKsQFbpVyNUbKsHj4xfL3akMs9IYf4N6CPG0OKK1CfZSwey5p/CdH6D10K+nvtMerXp+L1rPfiGEsFZvAY51h9XN9TPGSPD4ZEgPTmENEtGqkB3Rhnd6GsMs/Y02pFA9dZceh1aVfZq8PcyTEMK6uQGd/Mr5EdQsKNdvrASPQuRdRhU0Poug9Tf1EEKI3KAp4F2+S/k8MU+1zMQthBBCCPFoXtLpdZTtWDZPbKwEj0IIIYQQ2WcA6nqX8Ma9iLsEj0IIIYQQ4qEKAIUL1y2cJ6qsJXgUQgghhHg0QYCtb7BvntlgCR6FEEIIIbKvBIBPKR8JHoUQQgghRKb8AJx8nSR4FEIIIYQQmXIFsHOxk+BRCCGEEEIICR6FEEIIIR6fWICk20kSPAohhBBCiExFAsRFxUnwKIQQQgghMnUGIOp4VJ7ZYNtM3nOQY0IIIZ44IxAn2SCE1QaPqZHHIvUSPEITYL0cE0II8cTtB6pINghhlS4DYRd3XgwwphjzxCwztpklqB1SjvLlismhIXK823cSWLR4E2WCgqhXo4ZkiLAKy9atI+bmTckIIaxXMrAt5kxMwM2LN/Es5inBY/u2DXnrzfZyaIgc79KlSBYv+Y3natXi27FjJUOEVdi2Z48Ej0JYv5+NqcbuR5cepe6wurl+Y6XDjBBCCCHEo/kFiDy88DDGZKMEj0IIIYQQ4qFigSURRyM4tf6UBI9CCCGEECJTXwPxO8ftxJiau0sfJXgUQgghhHh0p4EFl/68xMG5ByV4FEIIIYQQmfoEiPz9w9+5deWWBI8i5zGZTJhMJskIIYQQImcIB96/E3HHtKrbqlxbfS3Bo5WJj09k8jfLCak3kHwFu+FboDuVqvVlxKg5XI+5JRkkhBBCPFtzgAXnN5/nt/d+g1xYxiPBoxU5d+4KIfUGsXgpVKu+nNcH/sugQRcIbfQbW7cHUL3WIPbuOy4ZJYQQQjxbrwP7dn+1m+3jtue6AFKCRytxM/YOrdp9Quky42jadDLe3iX/e8/NrTD1639CaKPFdOwygXPnr0iGCSGEEM/OHaAlJk5s+WgLWz7ZkquqsCV4tBJfTV6Gh2d7Spdu9cA0BQvWpGKlj/hkxGzJMCuWmJjIrdu3LX7ciYuTTBNCiJznGtDEZDQd2DZmG6u7ryYhJiFXbJit7NucLyU1lcVLt9Ky5e5M05Yv34XvZ48lJuYWnp6uknlWaMSkSaxatQpH28xPTxPg4OXFnz/9JBknhMhMIGDpheEyEA14AhUeEBidAVIlWx/qEtAYE/MPLzrc7Or+qzSf1pyA+gGgk+BRPMlbl2vXSUryw8nJN9O0er0dXt5VOHb8AnVql5fMs0J34uL4sn59mgcFZZo22WikwuLFkmlCCEt8CTRN95qjCgCT0r3+NjAdqApsUoFimHrPCQgCzgJvAn9K1j5UNNAaeC/qeNSHC55f4FihawXqDquLd0lvq9wgqba2AqkpqdjaOqHTWXabYmfnyp24BMk4IYQQ5noCRcweaSUMW9K9XgSYm+6z3wKh6lELKA2cA9YDZSRrM5UEjAbqGJONWw/MPWD6rtJ3rOm9hos7LmJMsa72kFLyaAW8vNxIiA8jNTURvd4+0/SRkYcpXqyRZJwQQghzsemep6i/yUBUFpcVCQxEK5EcAAyW7LXIfqAx0CI5PvndA3MOhBycd1DnFeRF8SbFKda4GH7l/XAr6IbeXi/Bo8g+V1cnalTPz9mzmyhZ8qWHpo2IOIKfbywBRf0l44QQQjzpYPQ24CNZkSUpwCpgLVDdZDR1iz4Z3Tr6ZHT+vd/sxcbWBjsXO1z8XUi6nSTBo8i+Dz/oRsvWH1C4cAiOjhm3kUhJieeXX95g4vgu6PXSIkGIXMgJmAe4ZPBePFo14jpgGzlnZDlHoBVaM6k1KtjIinaAPVrVqoxDlrO0AAoAv0pWZEsqWnvRP4G3gGCgnjHFWD7hRkJQwo2EgoC7Ov5Tcm3weOdOAo2bDiEpKTlLn1u66DNKFC/43/OQegM4e1b7jbC11fPH71MoGVT4ocsYOXou30xbSYniBdmxdSo2NrkreKpapRSffPQin41uRKtWC/Dzu7czzI0bF/jpp/50bF+Q1i/Xk1NS5BmvDBrE1cjIR15O9YoVGffhh/89/+Lbb/ni22//e/7e66/z7oABD13GkRMnCO3QAYCNixZRuVy5J/Gb3RjwAJargDHtdV+gH1pHh81AN+BqDthFHmgzbtgDxbIRPE5T2/aiBI/PVDm0Th+om5d66hibCyyU7HlkycBB9UijM3vkqEaRjzV4TE1N5cDB0yQkJKHTYWEHDx0JCYn3vHL9+i0io27893zs+B+YM2tYpoFrZNQNvLzccu2R9Wq/lhQLzM+Qoa1JTChGwUIh2NjYcvXq36QkH2LEx13p2rmJxR1rhMgNdv39N+cvXnzk5RjSDY0UFxdHZHT0f88nTp9On86d8XR3f+AyUlJT//tMSuoTH8HkDbThUsy5AMOAD4AN6gL/rOctTQb+AuyAxGx8fj/acDE35Wh/puoDBc2eXwJaAr+RKyfgyxFMOTVvn1i19QvP1+SriYMsShsYUCDD14OCCnP6dBiLlvzG+0O7UrpUkTx/JDVpXJ1Df8/mwMEzHD5yDqMxmRIlmlCrxmAMBmmFIPKeuV99RVxCwj236m6urv/VPjTt0oUbsbFUCA5mxoQJWmCXksKtO3fuWY6vl1eGy/fy8MDGxobwyEimzZ3L8ME5ul/AbeAjIB/QB3gVmPiM1ykKqPsolxM5ynOEhcAnkg3iiQaPbq5OlC5V9JGW0aJ5bX7dtJcjR88zasw8Fs7/WPYYYGNjQ5XKJalSuaRkhsjznqtV6+E/cqpE0dXZmZqVK2d5+a4uLvTs0IHPJk1i8uzZvNqtGz4PCDRzUGnFHBU8vpQDgkchRG6LQ3LyytnZGXh7cEcAli3fzLHjF2SPCSGeugHdu+Ph5kZkdDRT5861hlVOG5LF2cLrQEZtXdzQqostuU64AF5os5dktd2MvfoeT8CQze11zebn7eToFiLrcnw9Z+dOjRk74QdOn7nEmLELpPRR5AnrTp7kXExMpulSjUbJrKfA39eX/q+8woRp05g6dy4De/bM6aWPz6m/B9K9vhat08kAIAQYhNYRIhRtFgzQpqIbA9RRwdhJ4DO0Xtzp1UCrJq+N1qs6ATiE1sllJXenrpugrjfvo7WBBPAGRqL12PVCKzGNQOvsM5K7s5m8DZRCa8t5Pd331wWGq/WwR5tS73tgCnc7E6X5Ca3NZTegL9AVyA+cB75Q7wshckPw6OBgx/tDu9L31fEsW76Zj4Z1p0yZorLnRK7VtlkzDhQtalHXOh0wrFkzybSnYHDfvsxYuPC/0sdP3347p65qNbS2afEqiDNXCW3oj0HAJJXmb7Mgr64KElOAWUAM2lA5y4FOaGPTpWmHNmxQigpKz6pgrIEK4H7lbieXzmilfMPNgsclKsidrdbBA6gCtAVmmAWPzdQyR6cLHnuizXoSodY1Cq1TxxjgebSeweaDYldG67G6XK3nHiBOrUMN9XdfHjusk9S+OvyQNFdUmn/kV0BYTfAI8EqXJoz7YiFnzlxi9Nj5UvoocrX6ISHUDwmRjMhhCuTLR/+uXZkwbRrfzJmTE0ofG6vgLo070AjoiFbV3B+tp3J65dBKCmcDI7jbY9sVbdiVeBWsnVKvTwJWAOOAn1XAEQDMBG6o9Thutnw7oAkP7x1dRH1uPvB6uvf8ub+EMb3KwFS0Us7m3J0dZSJaW88ZaPM490v3uUJqXWuhlULaoJU6vg10sTB4tAUWqDzIjmQVTF/OAYd1nMqvhzlmQZpnxYDWkafwI+yLDtw/aoHIDcGjvb0dHw/rTo8+n/Pjii0MH9aN4DIBsveEEE/VkH79mL14MVHXr+eE0scFGbx2C23olFFopXkZ8VIX3Ne5d+y4NkBx9fops9cTgY/V8iqjldgNQispfDdd4IgKLn/OZN3tzAJFW+4dANmSC/l7aNXUg7h3Wj2TCoo7olVPTwBOp/vsOO4OGWQEflTBY5CF+Z6CVo2f3XHhUoBwOZsei2TgU3Xjk919ESnZmIOCx+sxt9j314mH/3rY2VKxQgmLltels1b6ePzEv4waM4/FP3z6WNc3MTGJfX+d4Nz5Kzg7O1KlckkCA/LLESKE+I+/nx8De/Zk5Fdf8fX33/N6jx74ens/q9Vpw70li0a0dot3LPjsCO4fdLiR+lsOrYo4/bUiCSiLVjr3ggrAsttO8KxazvPAXmC9er6Du20vH8QJaIg2zuCBDN43qeC1sVq+efCYyv2lsfHZuB6eeIr7uSda+9C8FBDO5P42qw9y/AmuiwNQXn757nPsiQWPv/3+F7/9/tdD0xQs6MvFsz9aNBuMra2e4cO680qPUSxf+QdD93d+LEPVGI0mpk5fycQvV+LsUh0vryBSUiK4fHk1wWX0fDlhIOXKBsqhIoQAYFDv3kydO5fomBgmzZjB2GHDntWqXAMuZOfeHjiTweuF0UpiGpLxbBZn0DrEOKG1GYwm+wN3m9DaNi5Dq0JOG0PpFlop5xQePDiyK9qMM/9wt/1kemmlkekbyMepINiaFFH5nVckAvocsi6F1c2NuFe1JxY8FizoS41qZR6aRpsNxvJRHTp1CGX02PmcOPEvY8bOZ8Wy0Y+0jqmpRnr0HsP+g+6077AXFxf/u79sJiOnT6/nheaDWTh/MA2eqyyHixACXy8vBvbqxchJk5g2bx5D+vXDz8fHmjbhQSWTN9WFuw5aW8YHBX1pM8U4orU5S8jmeoSp7yqv/tZGmwP7K+Ai93bOueenWwWNHmhtFjMKdNOqlJOfUB46qiA6O5K5tyNPZkbKWfdQTmp/PIl9EY3WREHc6+oTCx7rhJRj6aLHm+d6vZ6PP+xO1+6jWPfTTv7+5xRVq2S/9PF/X//IoSM+tGv3AzrdvaWfOp0NJUu+hLt7EXr2bs2+3f/D19dDDhkh8oqHTPP5Rq9eTJ0zh+iYGL6aMYOxZnNiW7HdwMtoQ/isf0i6JLSq3+Yq8Nv1CN9p5O58vtNUAPkHWlXtg4LHG2iloEFoc2VnVIpaT/3d8wTyyRbYiOVtJDMKfmurANkSBhVUWzou8xoV3LdHG/vSXILaX2d4fNPe+aCVIj8sAFuOVprYJYOg+w6wLQv5kT5vNqnjIDtS0Eq+H9R56TpaEw+RwUlgVdq3bciYsQs4dvwCn4+bz/Klo8nOVM63b8cz+ev1dOq0977A0Vy+fBUoEdSPr6euYOSIPnLECCHw8fJiUK9efDZpElPnzePtV199lm0fH5fFwIeqpGU798+J7aGuGVFonVKaoVUxt+T+Er5CaCU6DyrV0avAJird6yfQ2rrZZXLBnwn8D62zRE/uDjUE2rA7bVSAtOkJ5FMKd8fRfBocgLHcOwB62qDsGbVx3ayCx0/VflioXrdRQVbaGJyvk/m85+4qzcNGDisKTFeBVkadnU6o4NGA1oEJsxsDPVBS7c9ZaB2wsjL/eTJaqbV4ymysbYUNBluGD+sGwLqfd/HP/pPZWs6OXYfx9KyLk1Pm1U3ly3fh5w37MBpl7nch8gzTw8/3Qb164e3pya3bt5n03Xe5YYsvquCxMvAL2sDh3mgdaIagtTFMmwtyLbAIaIpW0hWqgoBGaEP7HOb+9obmSqANATMObVigEmhjPE5Gq3LemMm6fqeCpK7AUhXMlUXrfb1eBZMDsLzTRU52S+VPUbPH++q999K9XpR7ew9HqyDxdZUfz6ONhdkU+NqC756u8tUSi1Ta9I/0pZLnzdbpVbQ2ts2BXkgVsQSPT1K7Ng0oX64YyckpjB3/AyZT1oO6s+cuky9fBYvSursXJiIyiZSUVDlihMgrMqnS8PHy4o3evQGYNn8+EVFRuWGrp6kLegngd7SSwcNoA2/v5e5A0aloYyj+TwUjv6PNRPOben05dwf5zki8WtZgYAtaj+i/0cY/XKaCw4dJQCtdnKUCj63AERUQRaDN6f27HMQZ+gtt0O/OZD7EjZ7sTxmZFZvRxhLt9ZS+TzwiW2tcaTs7A58M70n7zp+w9qed/LP/VJaXYW9nIDk5zqK0RmMKNjap2aoeF0JYKQtuSgf37cvUuXOJjI7mq5kz6diy5ZNeq9toM8joyV4bsfqZbTValfQKtNI8P7TqyL1oQ+OkDwCHoHVwCVGByDW0IXfCH/C9aZ1rLqKVfhVGK+n0Q2tLeVAFq+bVpD3ROkRcTbfMm2gDoY9Hq6p2UkHoHjKu+qynCkzSl0aeQJv+8E4eOroT1fU/JxUgJeXAdRK5KXgEaNmyLpUrBbH/wGnGjv+BYoEFsvT5ypWCGD9xmfqtfHhUeO3aQUqX9MXW1laOGCHEfzzc3Hirb1+Gjx/Pt/Pn0+DJzwxkRBsjMbvOW5juBlrVtCUuWhDInnvA62E8vISSDILW9M5amCfnHhJIncpDh20xtBK+X8har+/MOJDxeJTXybxzTgW06S5XkbU2j+IZsdoI385gy7D377Z9PHrsfJY+X7FiCdxcw7h8+a9Mb8R3757EK12fl5JHIcR9BvbqhY+XFzdjY5m+YIFkiMhR9zfAMPX4CJiDVjL8L9CXx9fjGqAH2rij5o/jaJ17zBU0W6dP0Dr07EDrBf627DIJHp+4l1topY9JScn88mvWxvG01euZOOE1NmzoTWzsg6cY3bdvGn6+p+nUIVSOFiHEfdxdXRnST5tCee2vv0qGiJzECW0KSH+0gdWvoU3bGMq9zQAKoXVcSv/wBapm8Hot7q+ym49Wqmn+CEZramHO1WydvNFKw1uh9dq/IbvMOlh1PaydnVb62KHzJxiNxix/vmH9ykya0I633mlA9RrDCA5uh729G2AiMvI4u3Z9gbvrQZYtHoHBIFXWQoiMDezVi69mziTq+nXJDJGTxKN1SspMWbRxIdMrijYkU/opI2PRpoZMSPddlswTHW7hOom8Ejw6OTmwduVYjEYT/vm8sr2cWd+9T1xcAoGBmc/I1KplXTb9MonUFC14dHZ2QJeF+uUO7UOpWrUUX3y5hKVLR3Hzpi16fSpBJVx5rf8L9Ow+ETs76fwlhLVaOn06ycnJeLi7Z+lzXdu2pVbVqjg4OGSa1t3VlU1LlhAeeffaWbJYMcl8YS02kvHwSMvQhlP6R7LoqdChDXlVVv0tiDbWpgFtLMw+aE0NbgJX0DqIHUNrHvBUh4N5rMGjra2eJo2rP/Jy6tWtYHFag8GWxqHVsv1dZ89dYcWqbdy6HUeN6sUJDPCnc8fGlC9XLEtBqBAiZwqtk70xhEsEBFAiIMDi9JXKlpXMFkJklT3wAtARbcxLfxVEotPpMGFKa5m6WUuj0+t0mA9RaEIbUusPtFES1qHN4W49waM1SUxM5uNPZ7Fg4QHKle9JoUKdsbGx4cChv5j/w2RealaEryYOwsnJQQ5tIYQQQjxOvmjV9/0AP51eh1dxLwJDA8lfJT/epbxxK+zGofmH+GPEH/99KCTkbWrUeIPY2Etcv36Kq1f/1l24sNU3KupEe6MxpT0QgzaO52QyH8lAgsesSE010rvfOE6dKUGfvvuxtbX/773ChetQrdrr/PHHJ7Tv9Ckrl4/GXqqthRBC5Bx70WaX2feQNBO5v6fzk3BJrUtmVdspwCjubSeZFzmj9Sp/F3DzCPCgYo+KlOtcDp9S9894Z3C6N/6wtXXAw6MoHh5FKVKkDpUq9QLgxo0LHDmyhIMH53lGRZ18B0yvA1NVnsc+7o3Ik8HjwsWb+OeAE+3bT8HG5v4s0OsNhIaOYd26vkz7dhVDBneQn6o8rFWvXly9ds2itEaTia5t2/KW6n0rhBBPwCH1eBAT2rA8T0M48IUF6VLQZjDKy+qhzZ5UxivIi3rD6lG2U1kMjo9eQOXhEUDduh8QEvIOx48vZ8eOcY7h4YfeResI9RqwQYLHR2Aymfh66ipCQ5dnGDim0elsaNBgJDNmN2Lga62l00weFvbvv8xp2BA3e/tM0/527hynwsMl04QQucEs4IxkwyPTo5U2jjQ4GRxC3gmhznt1sHOxe/xfpDdQrlxnSpduzd6937Bt28iiiYm31gATgJFoM/lI8JhV165dJyLSER+fkpmmdXMrREJCfs6fv0qpUkXk8M+jdDodRdzd8bCg162fs3OemqpCCJGrycClj84Orf3hAO8gb1rNa0WhkEJP/EttbR2oXftdSpRoyqpV3QzXrh0YDgSh9di+/ajLz3NzSN6+E4/B4INOp7covYdHINfCZew2IYQQQmQthgPmAgMCGgbQe2fvpxI4mvPzK0evXtsoXboVQAdgOY+hLWyeCx6dHO1JSYnBZLJsUPGbNy/i4+Mhp4AQQgghshJfTQE6BzUPovPazjj5Oj2TFbGzc6Vt28VUrNgdtGGBZqGNHSnBo6X883vj4X6TmJjM58K+ffsatrZhlCheQE4DIYQQQlhqEDCgaL2itFvS7om0b8wKW1sHXnrpO/MSyBESPGaB3saGga+15I8/Ps209HHHjrH07NYAe3s7OQ2EEEIIYYkawDj3ou669ivaP/PA0TyAbNVqHvnyVdABQ9GmnpTg0VI9ur1I0SLn2bz5owwDSJPJyM6dE7C3285bb8owPUIIIYSwiD0ww8Zg49jmhzY4+zrnrJWzd6NNmx+wt3czAN8AbhI8WsjOzpalC0cQVPwgs2bVYP/+2URGHiUy8hiHDi1g3rwGeLhtYM2KMTg52cupIIQQQghLDAQq1nyjJkXq5sxRWvz8ylOv3ocAgcCw7Cwjz05P6OTkwPczP2D/gdMsXfY7R4/NxmSCUqWK8M7gTtSpXV5OASGEEEJYyh0Y6lrQlfqf1s/RK1qz5mAOHfqBiIgjg9BKIC9L8JgFlSsFUblSkBzyQgghhHgU/QH/OkPrYO+Ws2stbW0dqFv3Q1au7OKCNsf2exI8CvEYpRqNbPv3X1ztMm/0fDgiAjw9JdOesQNHjhBz86bF6X29vSlXurRknBAiuwxAX2c/Zyr2rGgVKxwc3IatW0sRHX2yN1rv6zgJHoV4TBo89xyLIyMhPj7zxK6uvBQcLJn2jL07ahROERE4WxDwxyQk4FyyJCtmzpSME0JkVwgQVLZDWRzcHaxihfV6eypU6MqWLZ94A02BlRI8ilzn1q04bt+JJzk5hcjIG8TE3PrveUpKKtevx2Iymfg3LIx1mzYB4OLsjI1Oh7ubG24uLri4uOBob4+joyO2estmGfrqs88k863QlBdfJMAj8wH+D1y7xqiwMMkwIcSjaAPoKrxSwapWuly5zmzdOhKjMaWtBI/C6iQlJXPlajTXrl3n7LnLhIVFcO78Zc6dv8rlK1EkJiRx8+Ztbt9JIDk5BZPJ9MBlbdiyhQ1btmT4nqODg/ZwdMTT3R1vT0+KFCxI0UKFKF60KEULFaJIwYL4ennh5uoqO0YIIURmdEAD1/yu5KuQz6pW3MurBF5eJYiKOhGqtsMkwaPIkcLDr3Pu/FUOHznH/gOnOHrsPKdOXyI6+iYpKan3pLWxtcHe3R5nP2fsvVxw9/DBvag7di52GBwN6O302LvZY+dqh8loYsOgDQQGNqJy5d4kJNzEZEolMTEWozGF+PjrxMVFqUc0VyJucD7sLNv37MNoTPnvO/V6Pa7OzhQuWJCKwcGUL12a8qVLUyYoiEL582NrK6eNEEKI/+QDShUKKYSto/VdHwICGhIVdcIfKAGcluBR5AgnT13kzz+PsuvPo+zcdZhz568QH594N0A02OBVzIvAqsXwKe2DWyE3PEt44l7IHfei7ujt9dg52Wn3RA8ReymWDW9swMurBOXKdc50vUymVFJSEklKukNs7EViYi5w8+YFbtw4T1TUKS5eOcmRE6v+G0jeYDDg5+NDzUqVqF29OrWrVaNCmTI4OznJThZCiLyrGODgX8nfKlfe379S2n/LSvAonpmwSxFs3vIP6zfsZsfOw4RHXCc1VQVgjgb8q/hToGoBfMv64lfWD7/yftg526HT657qeup0egwGJwwGJ5ydfcmfv+o976emJpOQcJ2IiKOEhx8iPPwQV6/+zeqNm1i5YQM6nQ5XFxcqBQcTWrcuTRs2pEr58hikZFIIIfKSIADvUt5WufI+PqXu2Q4JHsVTkZpqZMfOQ2zY+Ce/bNzD0WMX/qt+dvZ1pkTzIIrULUKRukXIVzEfBieDVWyXXm/A2TkfgYH5CAwMBbSpKxMSYrh06U/CwnYTFraT3f/sZduePXw2aRL5fHxo/NxzvNCgAS/Ur4+vt7ccIEIIkbsVAHAr5GaVK+/qWhCdzgaTyVhAgkfxRBmNRrZtP8jSHzezavU2wiNiAK0KunDtwpR4sQRBzYLwK+eHzkaXa7Zbp7PB0dGboKDmBAU1ByA5OZ6wsB2cObORM2fWs3DlKn5YsQI7g4H6ISG0f+klWr/4Ij5eXnLgCCFE7uMO4ODhYJUrb2/vlhY8ukvwKJ5IwLjvrxMsXLyJlau3cflyJABOPk5U7l2ZoOZBBDYKtJoxrh4Xg8GRYsWaUKxYE55/fiIxMec5e/YXjh9fxZZd29i0bRsDhw/n+eeeo0OLFrRq2lR6cgshRO6ht+qV19uTaacCCR5FVkVExLBk2e98P/dnDh46C4CDpwMVulWgXKdyBIYGYusgh1IaT89AqlV7jWrVXuPOnXBOnFjD0aNL+eWP7fz8++94uLnRvkUL+nbpQtXy5dHr9ZJpQghhveIBUhNTrXLlk5Ji0zqGJkvwKB5JaqqRnbsOM2PWWtas28Ht2/HYOthSskVJqvSpQmDjQOyc7SSjMuHsnI+qVftTtWp/YmPDOHZsBQcOzGHWosXMWrSIyuXK0atjRzq3aoV3HpzWMCExkV1//ZWlz1QMDs6TeSWEyLGuA8THxFvlysfHxwDG/7ZDgkeR9Yt5QhKr127nq8nL2PfXCUyY8CruRcMetajYrSLuRd0lk7LJza0wtWq9RY0ab3Dp0m727/+eYydW8cZHH/HpxIn07NCBQb17E1i4cJ7Jk8joaLr36UPLUqUsSr8zLIzxX35J04YNM03bY/VqHC3o+X47KYl85cvLASqEyK5zANGnogkMDbS6lY+JOZs28cZ5CR5F1g6eG7eYM3c906av5uy5y9gYbCjRrAQ136hJQMMA9HZStfq42NjoKVKkLkWK1KVJkwkcPbqUvXu/YdKMGXw7fz6tmjbl7f79qVK+PDY2Nrk+P4K8vZnWvLlFad/65ReL0k39/HNib92yeB08LZjGUAghHuA0YIo+GW2VvUOjok6k/feEBI/CwjuOW0ybvorJ3ywnMvIGdi52VO5TmZpv1NSmWdJJHmVHcnI8R48uxWjMvA2MXu9AjRpvEB9/nbNnN7JkzRp+/OknQuvU4bN33qFmlSrodLIjsqJU8eKSCUKIpyUMiLq0+5KvyWSyut/rsLDdACnAEQkexUPduhXHt9+t5suvlhAReQNHL0fqfViPaq9Vs9qxqnKSxMRY9mx5lzeqVbToh2Tz+fM4lBpCz55buXTpT3bt+oLftv/E5p07aRYayvDBg6lRqZJkrBBC5Dy3gb3hh8Obx0fH4+RjPbOOpaQkEBa2E+AA0uZRPEhCQhLffreaiV8t4cqVKOzd7Kn3YT1qDallVQe8NfBwcOC9OnWwtaDqOTElhb1o40gWLlybDh1WcvXq32zbNpp1m9axYcsWWj7/PJ8MGUKFMmUkc4UQImf5NTkuufnp9aep2L2i1az0hQtbVYcZNmXlczayv/MGk8nEup92UqVGH94e+g1RsbHUea8Og04OInRMqASOOYxOp6NAgWp07LiKnj23UbRoI1au30DN5s158+OPibp+XTJJCCFyjlVA/KEfDoHJelb60KH5gMkELJTgUdzj6NHzvNhiKC+3/ZCTp8Oo0qcKA48PpPH4xrj4u0gG5fAgskiROnTpsp5u3X7D06ssX3//PWXq12fq3LmkpKRIJgkhxLMXBvx+YcsFwg+HW8UK37x5kZMn1wDsA45J8CgAiL0Vx5tDJlOpem82/rqXIvWK0HdfX1rMaiHtGq0wiAwMbEjfvnto0WIGCUkGBg0fTuUXXmDLrl2SQUII8exNNKYYjTvH77SKld29+0uSku4ATCCL5aUSPOZCaVXU5Sp25+upK3At7EbbJW3p8UcP8lfOLxlkxWxsbKlSpR8DB54gJOQdTpy5QOOOHek/dCjXb9yQDBJCiGdnG7D16NKjXNp9KUevaGTkMf75ZxZoPazXZvlaJPs6d7l6NZou3UbyctsPuRIRTe13a/Pa4dco17GcDPeSizg4ePD88xPp128fhQrVZuaiRVRs0oTlP/+M0WiUDBJCiKfPBLxnTDUmrR+4ntSknDldodGYwsaNb5OcHGcE3iEL0xJK8JjLGI1GFiz8lQpVe7Jk2e/kr5qf3jt60+SLJhicDZJBuZSfXzl69NhM06aTiboeR4dXX6XjgAFcCQ+XzBFCiKfvL2Da1f1X2fLJlhy5gnv2TOHs2V8BlgC/ZWcZMlRPLhAdHcvgtyezaMlvGJwMNB7fmJqDa2JrL7s3L7CxMVCz5puULNmCX34ZzIr1P7H7n3/4bvx4moWG5vgS58Ph4TT94QeL0p6Ijqap7HIhRM72EdBg18RdlQpUK0Bwu+Acs2Lnz29my5aPAdMZ4E3UpNYSPOYxW/7YT98B4zl37goFqhXg5Tkv41fOTzImD/L0DKRTp1Xs3/89mzYNpVXv3rzeowdj3n8fF2fnHLnO+f38+HNTloYXo4C/v+xsIUROdgfobEo17VjTa423k48TAQ0CnvlKXb36Dz/+2IHk5LhbQHcgOrvLkuDRSiUmJjHq8/l88eUiUkxGag+tTYNPG0gVdR6n0+mpUqUfhQvXZe3aXkyZPZutu3czb/JkKgYH57j1tbW1pURgoOw4IURucwJonXQ7af2yNstc2i5pS/Hnn920qZcu/cnSpa2Jj49OVIHj7kf67Zb9a32uXoumW8/R/L75b9wKu9FyVstnelCKjII4HdfiU6g2e4FF6SPj4qhc5/Gdjr6+ZejefQvbt49m166J1Gvdmmmff84rbdvKzhFCiKdjO9AhPiZ+ydLWS92afdOMij0qorN5uk2JTpxYxZo1vUhIuBkH9ANWP/KNv+xb67Jn7zE6dvmUfy+GU+LFErSa2wpnP2fJmBzGycmXAQNPZe1ktHV4rOtgMDgSGjqGokXrs3p1d3q+9Rb7Dh5kwkcfYW9nJztJCCGevA1A0+S45BVr+67N/+/2f3lh0gs4eDg88S9OTo5jy5ZP2LPnfxiNqTFAF+CXx7Fs6W1tRb6buYaGjd/k4qUInvvoOTqv6yyBYw6l0+mwt3fL0kOvfzIBXfHiz9Onz5/456/OlNmzeaFLF65cuyY7SQghno7dQA2T0bTlwJwDTK84ncOLDmNMfVLDqpk4fXo906dXYvfuLzEaU/8GQh5X4CjB4xNkMpk4fOQsi5f+ztIfN3P02Hlt9shsSEpK4fU3JvHaoEkY7XV0WNGBhqMaYqOX3Scs4+ERQI8em6lSpR9bd++mVosW/HXwoGSMEEI8HZeApsA7Ny/evLnylZXMrDaTI0uOPLbxII3GZE6d+onvv6/D4sUtuH79dDzwKVAPOPk4N0aqrZ+Anzfs5v1hM4iLL0r+/FUBE5cvz8bbK4KxY/rxfOPqFi8rNvYOXbqN5OcNu/EN9qXjqo54l/SWTBZZZjA40aLFdxQsWJ316wfRuGNHFnz9NS2aNHlq63AhLIzWffpYPnyQycT0CROoUamS7EAhhLVLAiYBSzEx/NqBaz1XdF7h6OTrRNn2ZSnZoiSFaxfG3tUeLPqJNJGcHEdY2G7Ont3IkSOLiY29nPY9y4DPgDNPYkMkeHzMJk5awpSpf/PSS6vw8yt/z3tXr+6nT/9+fPj+ZV57tVWmy7p4MZzW7Yfzz/5TBDYMpP2K9jh6Okomi0ego0qVfri7F2XFis607duXSSNGMLBnz6cyHmRCYiJecXF837KlRemH/f47t+/ckd0mhMhNLgOvA2OBPnGRcV33TdtXfN+0fTq9QY9XSS98SvrgHuBOxOGIez7477/b+PXXd4iNDSMq6gTR0WdISYnXIkmtdHMR8B1w/klugASPj9HmLf8wZepeunbdjL29633v589fmc6df2HMuMZUqhhESK2yD1zW/gOnad1+OP9evEbFHhV56duXsHWU3SUej+LFn6dHjy0sXdqKwR9/zIWwMMZ++CEG2yd/jDna2lLUw8OitC7SsUcIkXuFASOAUUBVoH5qcmqDyKORZSOPRvoBdoDe/AMXL27n4sXtqWhTCkaiDQm0HdgM7CUbUw1K8PgMmUww6vP5NG48OcPAMY2Tkw8NG45nzNiRrFszLsOS6W3bD9Ku08dEXb9J/Y/qU//T+uj0Mi+1eLzy5atAz55b+fHH9nz53XdExcQwY8IE7AwyVqgQQjxFqSrw2wt8oV7zArwBd8AFrc3iKSAOuAnEqODxmZAeF4/J1WtRnD1nolChkEzTBgSEsv9gDDdibt333m+b/6Jlm2FE34il+bTmNPisgQSO4olxcyvMK69spFixxsxbtozub75JQkKCZIwQQjxb14HTaHNl/wFcBbYC+1QQGfksV06Cx8fkypUoXF2LY2OjzzStXm/AxaUYYZfubcuwYeMeWrX9kDsJCbSc1ZKq/ata2Gg25zEZTSTcTCAxNlFriSFyLHt7dzp1Wk2pUi+zdO1aOgwYIO0MhRBCPJBUWz8mBoMtKSmWl9gkJcXh4HC3PdeKVVvp2n0kRltoNa8VZTuWtcp8uBNxhz1T9nB67Xm45YqJVOz8EynbqQyV+1TGzkXasOXM49eZtm0XsW5df9ZtWkjrPn1YNXt2jp0TWwghhASPVi8woAAJCadITo7DYHB6aNqEhJvY6C5RuJAfAD+t30XXbiMx2etov6gdQS8FWWUeXNxxkZ96bqJi4CA6N34FFxd/TCYjN26cZ9/iqcxb8APtlrXGs5jnPZ9Ljksm8mgkt8Nv41bIDZ8yPtjay6H59ANIJ15++XsMBkd+2z6LNn37smr2bJydnCRzhBBC5JzgMTz8Okt/3MzefceJj0+kQAFfmr1YiyaNqmFrq7eajHRzc6JRwxIcPryQKlX6PTTtgQPf07xZBRwd7dn021907DICox46LGlPiWYlrPJAij4VzbpXfqf186soUKCa2Tt6vL1L0rTpZI4dq8vydu/Q/Y+u2LvZk3grkZ3jdnJ84QW8HCrh5lqIGzfPcyN1HRX7lqHWkFrYOkgQ+TTp9Xa89NJ0UlOT2bRtHt3efJOF33yDo4ODZI4QQgjgGbZ5NJlMfDdzDVVqDmHFanfsHD7GN99kwiM78s57u3gu9E1On7lkVZn52Se9OXx4HJcu/fnANBcubOHc2akMe/8Vtm0/SPtOH5OQnESbhW2sNnAE2PLRFupXG5cucLxXcHB7ijl3Zc+UPcRFxvFDk0WYttWhT6dDdO60lubNp9G1ywZ6ttnPzZUlWdxiCYm3EuUsfcp0Oj0tWsykbNkOrNqwgX5Dh5KcnCwZI4QQAniGJY9fTV7G1OnH6dp1D05Ovv+97u9fieDgdpw8uY6mL73N5l/HUrSIv1VkZuHCfvy4+AM6dOlMYGA/qlTph7Oztm23b1/jr7+mc+XyApYtHs7Vq9G0avchsbfjaD2/NaVbl7bag+hOxB0id6XSulfbTNPWqPEGCxcv4NLuy1Twfo/q1Qfel8bZ2Y9mzaaydetnbHxrLS1nt5Qz9SnT6w20ajWX5OR4Fq5cib2dHTMmTECv10vmCCGEBI9P37HjF5g4aTPde+zCwSGjwYJ1lCrVkqSkW7zx1hRWLx+DjY11dDuuVrU0+3Z/w/8mL+PHFTW4HmMHmPDzNdK+3XO89cY3xMTcom7DQdy4eZsWM1tQvmt5qz6Ioo5Hkd+nOnp95p1hXF0LoLvuS2qiD9W6vvbAdDqdjnr1PmTOvJ8JPxROvgr55Gx92j8Oto60bbuQxYtf5vslS8ifLx+jhg595JloouLj2XzesskPrty6JTtCCCEkeIRp01dTo+bQBwSOd5Ur14nZs//HqdNhlC5VxGoy1dvLjVGf9WXUZ32Ji0tAp9Ph6GgPwI0bt2jZZhiXr0QSOjqUyr0rW/1BlJqUiq2t5Z0qDAYnKlfqg0738FYTer0d5cu+wqmf1krw+IzY2bnSseNK5s8PZdzXX1O0YEH6de2a7eW5ubhQrFo1ZsXGWpa+bFny+fjIjhBCiLwePG7fcYTnX/g203Q6nZ6goJfYum2/VQWP5pyc7nY0iE9IpFPXzzh85BxV+lWh7gd1c8VB5F7EnajrhyxKm5x8h+vXz+Dvb1nQnD9/FfYdmSVn6jPk4OBBhw4rmDu3Pm9+/DGFCxakaYMG2VpWAX9/Fk2dKpkqhBBW7Jl0mImNNWJn52pZYOJelKiom1af0UajkTcGT2bjpr2UaFqCF79+EZ1N7pg5xrO4J8nu54mMPJpp2jNnfuHOnQhMphQL8y0FG1sZy/5Z8/AIpEOHlaBzpOvAgRw4elQyRQghJHh8etzcdCQlWVZtdePGBby93a0+o//39XJmz/kJ/0r+tFnUJleNY2hja0OdD2vxy8bBJCfHPTDdrVtX+WPncIKaBxEWtsuiZYeF7SJ/lfxypuYABQpUo1WrecTejqd9//5ERkdLpgghhASPT0f95ypw+vSGTNMZjamcPbuBhg2su13g5i3/8OFH3+Gcz5kOKzrg6OmY6w6kMm3LUKBdIst+bMv162fve//y5T0s/rEpIaNL03BkQ/Yfmklq6sOH4UlMvMXRUz9QqmUpOVNziFKlWhIaOoYzF/6l51tvkZySIpkihBB5zDMp/hrQ/2WavDie0qVfxsHB84Hpjh5dQpVKdgSVKGy1GXwxLJxuPUeTYjLSbk6r+2ZXyS10NjpCPw/lUPAhlo1riIexMvnzVyE1NZlLl3eR7HOOxnOeo1ijYmCCfE3/ZvPmj2nceCw63f3Dv6SmJrJhwxuU7p0v1+aZtapV6y2uXv2H9ZsXMeqrrxg5dKhkihBC5CHPpOQxuEwAbw9uwLJl7bhzJyKDFCZOnFjDwQOfMnnSm1YzTE96CQlJvNJjFFeuRVH/k/qUeLFErj6YdDY6KnavSN9/ulN3ji/u3ffj0+8YL6wuRa+d3bXAEUAHL3z5AjcKrGf58k5ERBwBTNqeNxm5evVvFi9piW3NI9QbXk/O0py2n3V6mjf/lnz5KvD511+zZuNGyRQhhMhDnlnDuyFvdcDZeS2jx9amePEuFCvWGEdHT6KiTnD48CK8PM+zYd0oAgL8rTZz3xv2Ldt3HKJM2zK5pme1RQeVgy2FQgpRKKTQA9MYnA20XdyGI4uP8NO0FiSct8fJwZ87CVdwLplCzfHVKd2qVq7pVJTb2Nu70a7dUmbPrk2vt99m788/UyIgQDJGCCEkeHxybHQ6BvR/mdYv12PZ8i3s+nMkUZEJBBTNR+8v6tCo0WAMttY7m8X6DbuZ+u1KPIt70nJWS3R6CYLuOwZsbajQrQLlu5Yn8WYiibcTcXAPwc7V7pEHohZPno9PaVq2nMmyZe3p8/bbbFq6FDuDQTJGCCEkeHyy8uXz4o2BbXljYNtck6kRETH0fXUC6HW0WdgGBw8HOdIeQmejw8HTAQdPySdrU6ZMW6pU6cu2PTMZ9803fDJkiGSKEELkcjKA3mNmNBrp3X8cV69FU/+T+hSqWUgyReRqL7wwCR+f0oyZMoUde/dKhgghhASPIiumz1jD+g27KVK3CPWGSWcPkfvZ2bnQuvV8jEYb+rz7Ljdu3pRMEUIICR6FJU6fucQHw7/DwdORVvNaSTtHkWcUKFCdBg1GcOrsWYaPH6/6zgshhJDgUTyQ0WjkzSGTuXUrjkZjG8nYhCLPqVVrCIUKhTBz4UL+2LVLMkQIISR4FA8zb8EvbPx1L4GhgVTuU1kyROQ5er0dzZtPBZ2BNz/6iNt37kimCCGEBI8iI1evRTNs+HcYnAw0+6YZNnrJVpE3+ftXJiTkHY6cPMn4qVMlQ4QQQoJHkZ7JZGLo+9MIj4jhuY+fw6eMj2SKyNPq1HkfP7+yTJoxgwNHj0qGCCGEBI/C3NZtB1i6bDP+lfypObimZIjI8+zsXGjadAoJicm8P2YMqUajZIoQQuQitpIF2ZecnML7H04n1WSkyYQm2DpIdgrNn39+xcmTay1O7+UVRIsWM3LN9gcGNiQ4uD2bti1h7caNtH7xRTkohBAic4WBkkCQ+r8H4AhMAE5I8JgL/LDwV/buO05w+2CKNS4mGSL+Ex19ivHVfAgplPkg8VFxcbTdsD+X5YCO0NDRnDr1Ex98/jlNGzbE0UFmEBJCiHT0QD2gHdAYKAZkNM/rfAkec4GYG7f46NNZGJwMNBrTCGRIR5GOr7Mzhd3dM01np9eTGw8gT89ihIQMYevWkXwzZw5DX3tNDgohhNC4Av2BgUAAoHPxdyGgQQD5q+bHp5QP7kXdObzoMDvH78xxK58jgkej0cS581c4ceJfHBzsqFC+OL6+nuhy8PV0wheLuHI1irof1MUryEtOAyEyEBLyLgcOzGP05Ml0a9sWfz8/yRQhRF5mAF4DPgG8nXycqPBKBcp3LU+BqgXuK0c4u/FsjtyIZxo8mkyw9qcdfDZqPlHRnvj4BGM0phB+bSY1qnvz+eh+lA0OzHGZdvVaNFOnr8K1gCt1h9WVU0GIB7C3d6VBgxGsWdOLCd9+y6RPP5VMEULkVZWA2UAV1wKuhLwTQtX+VbFzsbO6DXlmwaPJZGLkmLnMmXeKZs1+xM+vnNl7Rk6eXEvTl95nzswBNG5ULUdl2oSJi7h1K44mnzbB3s1eTgchHqJCha7s3v0l3y1YwLsDBlAgXz7JFCFEXmID9AW+1NvpXaq9Wo36n9XH0dPRqjfomVi9dgdz55+ma9df7wkcAXQ6G0qXbkXLlqvo038qly5F5JgMu3I1ihmz1uFe1J2q/avKKSFEZj8yNgbq1RtOXHw8o//3P8kQIUReYguMB6a7F3Z36bq+K02nNLXqwPGZBY+pqUZGjJxPs2bTMRicHpjO1zeYcuWH8MWXi3NMho3+fD5xcQnUea8O9q5S6iiEJYKD2+HvX4m5y5bx76VLkiFCiLxAD0wD3i1Yo6Cuz+4+BDYKzBUb9kyCx6PHznMnriC+vsGZpi1XrjPrNx4gJSX1mWfWxYvhfD/3Z9yLulOpVyU5LYSw9IfGxpb69UcQn5DAuP+zd9fhTZ5fA8e/SZq6u1Ck1Cm06HAoMmC4DNlwmDB/5xuzH2PMlbnAYAx3d3cYUKy4FKu7p0neP1I2rDRAC2lyPtfViy25kzzPee4kJ7fKtoVCCPOnAD4BnqjVrhaDVw7GKcDJfD7TH8SLnjhxAU/PuhizPImdnTtZWTbk5OQ/8GB9+/1cioo0tHyzJWo7tbw1hLgDYWHd8fWN4a+5c7mclCQBEUKYs5HAKwGNA+g/tz+2bua1zu0DSR6trFRotcXGp+8KHSor1QMNVEZGDpP+XIa9pz31htSTt4UQd/ozXKGkWbNXyMvP59dp0yQgQghzVQ/4xtHPUTFgwQCzSxzhAc22rls3iJSUxej1OhSK2+ev2dkX8fXR42D/YIP/x+RlZGbm0ua9Nlg7WMtbQ5TrZFoaLjblj4tNKyhAr9dbREwiI/uxfv07/DptGq889RROjo5SUYQQ5kQN/KJUKR17T+ltVl3VDzx5rB0UQPXAQhIStlCjRpvblt2792f69WmJSvXAJoZTVKThh5/mY2VnReNnG8tbQ5TLz68B355OgzMZRpUPqBFrEXGxsrKlceNnWLv2DabNn8+YoUOlsgghzMko4KEGTzYgqKP5blv8wNZ5/OLTZ+jV9ykGDlqDs3PgLcucOrWKtLTZvPDcdw80SPMXbOLc+UQajG6Ag7eDvDVEuRo0eIIGDZ6QQNwyNqPYuvVjJk6axKiBA7G2lpZ8IYRZcALedfBxUMR+aN4NAg+sOa9J4wh++G4IM2bEcujQdEpKCkvv0ZOfn8rGje+z759nmTfrfVxcHmzC9uMvC1EoFDR7uZm8NYS4R3Z2HkRHD+PYqVNs3rVLAiKEMBdDAf+mLzbF3sNeksfK0rtXazatG4eD3VT+nBzG5MnNmDSpMQsWNKJ1y/Ns3zyR8LDqDzRAh4+cZeeuI9RoUwOPcA95awhRAerXHwUo+PXvvyUYQghzYAU8Y+9hT6OnG1nEyT5QoSGBTJ08luJiDVeupGFto8bH2w2lUmkSAZo8ZTklJVpihsegUCjk7SFEBfD2rkNgYHOWr1vH5aQk2bJQCFHVNQbCI/pFmOXs6hspTeVArK3V1Kjhi5+vh8kkjnl5hcyYuRZ7T3si+kbIW0OICqJQKImJGUlefj4zFiyQgAghqro+gLLeYMtYyk8p17tsy1bs4EpiGlEDo7B2lEH9QlSkiIje2Nt7MmXuXDQlJRIQIUSV/T0MtHPwccCvoZ8kj5ZuytQVKBQK6o+qL8EQooLZ2roSEdGXI8eO8U9cnARECFFVeQERAU0CLGb3OUkey5CcnMHmrXF4RXnhXddbAiJEJahb9zF0epizdKkEQwhRVQUBdv6N/C3mhCV5LMPylTvJzS0gsm8kSpWESYjKEBDQBFfX6ixavZri4mIJiBCiKgoF8AixnBVZJCsqw8zZ61AoFYT3DpdgCFFJrKxsCQ3twZnz59l36JAERAhRFfkDOFd3luTRkiUmpbNlSxw+0T54R0mXtRCVKTKyHwAzFi2SYAghqiIXADtXO0keLdnyFTvILygiok8ECqWs7ShEZapWrSmurrVYsmYNGo1GAiKEqGoMa2ZbULogyeMtrFy1C4VCQVj3MAmGEJVMpbImOLgzZxMSiD91SgIihKhqCgG0xVpJHi1Vfn4hmzYfwKWGC56RnhIQIe6D4ODOAKxYv16CIYSoajIACjMKJXm0VPv2nyA5JZOg9kGo1Koyy+n1eorziinMLKSkSBY4FuJeBAa2wNraiVUbN6LX6yUgQoiq5AxA2sk0izlhK7nm11uxahcAtTvXvuX9Oo2OwzMPs/+POPJOK1BpHdDaZeLXzJVmrzTDt76vBFGIO2Rn505AQBN27ttGVnY2ri4uEhQhRFVxEtCnnUizmFGPkjxemxjq9KxavRu1g5oarWvcdH9RdhHzBs3H+nwMnVrOxadNNAqFkuLiXE6dWsnC7h/Q4I3qNHmuiUUNnBWiIgQHd+bs2XVs2rmTnp06SUCEEFVFApB+ccdFD71ej0Jh/gmAdFtfIyMjm4OHTuMX44eDt8N19+n1ehaNXERA/mD69p2Br299FApD+KytHYmM7MfQxzZx+Mssjs49KsG8Q8W5xaSdSCP1WCrFObJYtCUKCuqAQqFkw/btEgwhRFWSA/yTGJdIYbpljHuUlsdr7NoTj0ZTQmCLwJvuO73qNMVxgbR+bOy/SeON7O096NVjCjPebkNot1Cs7CS85SbsZzPY8tEWLq3PwJ6aKFCQpz+PT3MHWr/XGs8wmbRkKby8IrCz82D73r1Yyq93IYTZWKXJ0zx8es1pogZGmf3JSsvjNbZtN+xwcavk8eBfB2n60P+VmThe5e4ejJ9dW85tOicBLce5DeeYHruQwJTnGDXwEMOGbmDo0PWMfuwQoUVvMavDck4sOSGBshAqlQ3+/o2IO3qU3Lw8CYgQoiqZDxQe/OsgWMCcP0ker7F9x2GsbK241ebmifuSCAxsbtTz1KjRmiv/XJGA3kbG2QyWjdhAvy5LqV9/JNbWjv/ep1bbExU1iEG9V7HmmV0kH06WgFmIwMDmFBcXszcuToIhhKhKzgGbzqw5Q0p8iiSPlqKoSMPef47hWtMVJ3+nm+7Xa5WAgqNH57J06dNMn96NhQuHs3//ZAoLM68ra2VlZ1GLhd6NLeO30Dz6fXx86pZZxt09hHbNP2fTB5skYBaTPLYAYNvevRIMIURV841Wo9Xt+GqH2Z+oDMordSDuJLm5BQQ3vfWuMk6BdvzySzTVq7cmIqIP7u7B5ORc4uTJ5WzdOoGOHT8jPLw3AElJcfjUdpOglqE4t5hLGzLpPLB/uWVDQ7uxacpY8lPzsfe0l+DdhYsXd3L06Jw7ekxs7HjU6vu/T6u/fwPUaju279kjF04IUdWsBnbETY1r0fiZxvg18JPk0dwdPHQagICHAm6678q+K2Qd1dKnz2yqVXvo39u9vCIJCupIkybPM2dOf3S6EoKDu3DizEJadOpzX45bV6Ij8UAiyYeTUSgV+Mb44l3HG4XKdCcb5FzKwYGg67qqy6JSWePuGEnGmQxJHu9SUtJBArNX8VjdukaVH7N0KdrW7zyQ5NHGxgU3t2DijsqKBUKIKkcHvKHT6DaueG6F1YgtI0z6u1iSxwpw9Og5AHzq+Vx3u7ZYy7InV9Cj87TrEsdrubkF0b//XKZN68yJE8uIGOGPg49DpR6vXq8nfm48m97fhrOmHv7+jdDrdcRd3Emh81LafdyG4C7BJhlrvV6PUqkyurxCobKIAciVKdTDg55hxu3V/sqqVQ/0WH186nH48HRS0tLw8vCQiyeEqEq2AZMu7Ljw5OaPNtPmvTaSPJqzAwdPYWVrhXtt9+tuP7fxHM75jalRo9VtH+/qWpOoqIGc1U2m2wdDKv14t3y0hTN/lNC/2wY8PEKvuy8xMY6lT40i8+1MGj3dyORi7eTvRE7JaTSafNTq27cm6nQa0rKP4FrzEamkFsLHpx6HDv3N4ePHiW3eXAIihKhqXgdabf5wc0RA4wCTbci5FzJhBtDpdBw8eAoHL4ebWgzPrj9LWFhPjNkyJjS0G7YudljZVG5OfnrVaU7+WsBjg5bflDgC+PpG8/ig1ewdd45Luy+ZXLxtnG3wbm7H8eOLyy175sw6nOuV3LRouzDn5NHQvX5Quq6FEFVTFvCYrkSXNXfQXJP8HpbksQIkJqaTnpFzy32pi7KKcHauZtTzODn5UZBS+bujbHx/E106T7ztmEE7O3c6tv+azR9uNsmYt36nNZt3jSUj40yZZbKzL7J20yu0+aC1bPdoQby86qBQKDkYHy/BEEJUVQeAAUVZRYXTu03nwrYLkjyam+MnDBf1xvGOYGgly86+aNTz5OYmYudtXanHmnM5h5IED/z9y++ODgpqT+ruIopzTW+7P69IL9p935gZ8ztx7NhCtFrNv/fpdCWcOrWK6XM60fyTMAIaB0gltSDOzoHY2rpw9ORJCYYQoipbBQzNT8nP+7vr34ati81k/H6F9a8eOXqWFSt3cez4eWxtrakfE0r3bs3x9jL9JWvij50zfGlVc77pvlrtarFv6WKio4dQXvPX8eOLqdmjRqUea+bZTLw86xhVVqFQ4u4SQfbFbDzDTW+bv/Ce4bjVcmPj+2+z4c+38HAytDilZR/FKaqY7vPaENBEEkdLo1AocHUN4vyFCxIMIW54ewDGLjuhBwpK//Xg5saiYgzdq6JyzQHSi7KKZs0bNM/jwvMXiB0Xi7WjdZU+qXtOHrOy8vi/V79j7fpkwsMH4u3dnbyCIiZN2cr/xr/CG691Z8yTvVAqTbffMTEpHeCWCZZrLVcSUhezbt3beHiEUq1aczw9Q29KJDMzz3Hk3BSGD360ci+YrRUlJRlGly/W5Fb6GMx74VPPhwEL+pOXkkfGqQx0Oh3uwZ1x9HaUrmoL5uZWi5MnD5Kdk4Ozk5MERAgDX+CIkWWLgNDSf7eVPvbql4cScALigA+B9RLaSrUOaKEr0U3d+fXOJqdWnKLj5x0J6RqCQlE1v+juKavIyyukV7+xKJS9GTbsdVSq/zLpsLCe5OW9znffDycxcRIffjDKZINw7FgCCqXiukkZuYm5rH5lNamblcSEjsHBwYeMjLPs3fsLDg7edOnyHa6uNQHIyDjN3PmP0vbLxpU+scMzwpMracvRaouvi/ctE8fiXHJ0x3GuHmvyFdHBywEHL5kUIwxcXGqg1Wq5lJgoyaMQ/0kHet9w20DgaeAdYOs1t+uAfEAF2AFngavLFygAT+A1YEXpcyyQ8Faq40As8HrqsdRXZ/ac6VC9ZXVavN6C2p1ro1RVrVGE95Q8fvzZNIqKY3n44bHcqpnIwcGLvn1nMeWvdnTqeJCWLeqZZBCuJKahslbhGuQKQPalbKZ3nE2joLH0HD4alUr9b9m2bd/nyJHZTJvWmXbtPuTixZ2cuDyT9t+2JKJPRKUfq9peTc1HPImLm0qDBqNvW3b37olEDAipcpVSCDe3Wuh0Os4kJBAREnLfXvfEmTOUaI3fWtTDzQ0fT0+5YOJ+KQJu3K+1Sem/R25xH6XJ49VksuCa2xOAF4Aw4DNgUWkZUXnygQ+A6Xqd/r3zm88PSNiSYOVSw4W6j9UluEsw/o39Tbq38J6Tx+ycfKb+tY3HHt/L7foXra0dadPmQ7765iOTTR6PHTuPnbthiR29Vs+SUUtpFjaB6OihN5VVKFRERQ3CxsaVuQv70OGzdox+fAi2rrb37XhjP4xlSov38PauQ7VqzW5Z5syZNRw4/y2j5w6Xt6uocq626idcur9LXLTt04dG7u5YKcv/wZWQlUW7nj357J135IKJqkoP/AO0AhyAHAnJ/fmdCgwGPtTr9U9nnsscuGXCFt8tH2/BxtkGr0gvPMM8canhQuL+RPNKHvftO46LaxNsbV3LLVu9eiumT08mP78Qe3tbkwtCZlYuga2rA3Bh5wV0J6tTb/DtF/oOCelCWO0+OPqV3NfEEQwTe/ot6M78/v2oW/1p6tcfjaOjtyGpz77E7t3fcypzGoOWPXrfj02YHhsbZ344dIkfDk0z8hHuhl19HiAXF8P78XLi/f3gtFWpmNm3L/Zqdbllp8bFcbhqVYUwIPAWt2uBROAUoDHh41dh6PbLBnZfc7sT8BCQBBySd/wdUQMPY+hSlcTx/jsO/B+GRcVbo6dDUVZR64s7LkZc3HHRGcPYVJMcFHnXyWNySua/H/Dl1k61PQpcycnJN7nkMTExHZ1O9+/Mp5PLThIVNdyoQaxRUYM4tuy9+9JdfSO/Bn6M3D2Y3RPX8/e8H9GmOgJ61L4FRA2qw6gxw6v8bC5RMaKiBhIVNbBKHbOdnWGnp+TUVLmAFedZ4PnSZPHa7kllaWKWAswEJpQmk6bGHlgD7OG/rlqAoNLbZwKD5DKXyRqods13fxjwMhACdJXwPFAaDJNq1pX+vwLwA3wA99Jrd9AskkcXZwfy888bVVarLUKnz8bOzsbkrlhiUhp6vR4bF8OxpZ9MJ9g91KjHursHk3Yw7cF9wbrb0eb9NrR5vw2aAg0KhQIrW9lxUlR99vaGcYQZWbKSSCX4Hvj82o8SoBaGiRhPAH2B7sA+CZVZibwhAckBNpYm4sckPCZFD1wu/TNJd51p1K8fSnLy75SUFGJldfvWxMTEOIKDHHB2Nr3ZtAUFxej14FLdBQAbFxs0hbnG/VTQ5GHrYhotqWo7tbzdhNlQKq2xtXUhNT1dglHxcoAbB5OewtB6twyYB8wGooE8CZfZOI+hlVaIe/+MvtsHenu50qljbf7559fbltPpNGze/D+eGdPLJAOQmmZo2biaBAY2C+Ts2Q1GPfbs2Q1Ua1ZNapEQFUyhAFtbN1LS0iQY99cyYDlQG+go4RBC3Mo99XFO+PAJ2j38Ok5O/kRG9uXGcZ1abTFr1rxO44YaenZvaZIBuHQpGTBsQwgQ1iOM7R/O4qHsF3F2Lnt3k8LCLOKO/sbAnztLLRKVJjExjvz8FKPL29t74usbYw7pI/b2nmRkXZRK8AA+Fq+2EUgohBAVnjz6+LizZOE4ho74iKNH5xITMxxPz3B0uhISErayf99PdOvqz+efvI7KRNcazM0tNCSPpWMe7b3saf5uDAs/HsqjfWdjZ+dx02OKi/NYtHgEdZ8LxC3ITWqRqDQbNrxDkO4YzjbljxfOKSriJCE8/vhyszh3GxsXUlNkKNZ9Zgd0wDChZtct7q+NYUxkdQzrBM4GzpXxXLWBXkANDJN0jgOLuX4clxKIKX3NahjWMVxb+qeVyyGEGSaPAEG1/Nmw5isWL93G4iXfsGvnFWxsrKgfE8zYN0bTqGGYSW+/o9MZJh06+jr+e1vMyBiKcnbx5+ctad7kLYKDO+Hg4ENBQRpnz65n6/aPCBnpQovX20oNEpXuq06diDBiIerjaWl0W3nabM7b2toBnU5HYVERtjY2UhEqnz/wLRAOTOHm2Z2jgG8wzPxMxLDd3dvA4xi6u69SYNi55P3SZDSz9LvGCcNi1INKyyuA6UA/oBC4CDgDr5QmpcNKk0lxeyml1yq7jPv1GCbESDIuTCd5BFCrrejbuw19e7epcgG4upuEUv1fy6hCoaDpS02p3bE2u3+YyM6VYylK16J2VuLfzItui5rg38Bf9l4WohKpVNbodDo0Gs1dJY97Dhzg3EXju70taBvETsC1C/SqgWAMayXaA79j2HlEf02ZLsDPwI7SJPI8EAD8WppoRpQmMWDYKu8TDC2XL2DY+URZmpQ+y3+7nOgx7LgxrDSZzC5NNn/GsIDybuCrSjh/BfBI6bnejVxgtQklY3+W/pVFU3rNTVmL0h8vd/U1jmGyVy6iaiWPVVl+vuGHrdLq5m51rzpedP2xK3qtHm2JFqWVUrb6E+I+J495+fk4OTre8eO///NPCo8epZZr+RsZ5BYXsyXHYtZIDr/hi9oVw+4iqzC0Fu6+IXFUYmgxzAAexbAYNxj2Sh6MoVVrMPA14AZ8hKFlssc1CSXAXmAk1//sHs31a07mAe+Vvs4wDC2dFb1lnh2Grve7Xf4jG8M2gPnyLq0wj5T+gLkbWiBOkkdJHu+rtLTy15FTqBRYqWT9RCHuJ2trR/R6/b9DS+7GiJgYOgeX/510KSeH7StWUFhSwtNLlxq1PeGp9HSadq2Sayt/V5okXvUwhpY/ewxb1elvKB9V+vcTUFyaIF5VDGwBmpcmjz1L7//shsTxKv0Nz3+ri3uxNBGoVvodVVzRbQalSez9oMAwPlRlgW/hi3dw7cbKJ54kj0KYtczMc8THz7+jx9SvP9KorTzvhUarJTPzLDt2GN/TFxnZz+idou43pfL+r106b9IkNBrjd+ir5u9fFauwnuu7XFcCC4A+wAjgtxvKX52+P6b071Z2lP7boPTf3XfxXRSIYR1Cdwxd6VeTr6rMDlhRek6WJhaIN9FjexvD8Arxn2OlPyQleRSiMqSlHSfzyFeMql/fqPLf7tpFWFjPSk8ei7RafJXp9LU2bp3SaYcOkeYTbbLJ44PQrGFDSzxtPYa9dR8GPgSWcP32hFdbzT7G0Mp4y99Upf9e7QouNPK1awOvYmix9MEwQSYHcMQ8uiHzMezsIkxLEdLNfau6eke/9oQQd6i2mxtPGplozIu/fz++A52djT6uf65c+XfmginSaoulot0/p4EvgA+AccBT/Ne9fPyahHBFOc9zsvTfECNesz6GvXyzMUywWYdhIk4ukAxU1hR7W+BTDDO770Y28CYY/fZR3MF3rR7DBBAwtFre9Lag4rvxVRhm0Jf5VrzmNa8ek+aa47z2PG3LuK88bwGhd3n8JcD40rpjrC9L/8Rdsvjk0dPT9b+3rBDCZGg0eSgUCpQqlQTj/vgSw7I7I4CpwNarvzMwLBw+BMPs59wyEiQ9hlbLjzBMoPmhNJG4kRLDWMdnSxO4Dly/j7aae9j9zJiqhaEF1fouH19YxnmVJRjDZCRjZGCYeazAMAnEif9aca924W8uTbaPVlA8HgH+uM39y4Hhpf99NUGbBrx8Qzmv0rryPjDpDo9hzx0mfzcm3Fny9pXk8b6yszN8fui0OqkNQpgQrVaDUqnEwc5OgnF/5GLovp6HYfJLCwwtTkWlCcFvGJbxeeqaL2s10B7DWMcJpQnNX8BQ4HMMY8uudoc5Y1jjUYlh8k1h6X9HXpM8KoHeGGaAV9a+2lpg7n2MaxLwxg23DSo9z/9hWMroqqLSxNQG8Ci9r+81salZGtNtpUnfjnJe+00ME6Ru1yVpU5r4/QbML+P4r00Qr2AYL/gzcOKGHwWe3LrFtDxr5e0nyWPVCkBpq4a2WNZPFcK0kscilEolarVagnH/LMYwgeaR0iRxYuntkzF0Rb8KtMTQcqfCMJmmdmmCAoZWoBcAP+BFDJNwDmIYwxiFYSb2m6Vl/waexNDqNRxD62YIUIc7a9kzddnAnBtuuzpgemPpX1k0XD9rPQkYUJo0fg604vb9Zt0wrMNpzHi2Y6XXvjyTMCyj9L/SJNhcKTBs0WlK6/NlgmmMNrL45FFZuiRHXlIeQpiip5cuxcG6/B627MJCo7YxrCqKi3NRKpWyu0zF2Yahq3bPbcroMHRHJmBYC1KFoaVOh2Fc2trShK9uaWKzozRZXHtDstQdeAwYWJpc5mFo1ZrGf93hOzB0k79cmmzqMSy+PQxDF7kz188K1wC/cPN2iGmlt++1kOtYiGEB9j4Yxhje72QiB8Pkqa+BxuXUp6rMBsOqAT4mdEwjgBmSPJoAR0dDC3thZiFCmJqHH/6CgoJMo8o6ZCWQd+Bt8/mGLMzCxXJ2fbkfZpX+lec4t16SR89/+06Xm/tT/s4nYGiRm3OL298rI2m61fIqF7GsZVdsMOwGdAbjZ7VXtMnA8xgmWHWn4hdyN5lY+/l52Dzo3fPOnUtk6fLtYELrhVp88hhYzRuAomzZQlWYHg+PMKPL2tm5Yj47W+spKEgl0NdFKoGwZCr+mxVuhaFb/y0Mk3C6UbFTPXsDtW64LQ3D8k03jusqwNBtPQ3Deo7rzPUCBNXyZ+I3Lz3QY1iybPvV5NFkWHzy6O5ueF9Ky6MQJpQ66qGgIANPj2r39Dz/t2oV/9u0qdxyxVotGmdnCbwwNa34b8KKAkNL40agNbC/gl+rmJtn0uffJkGdh2HIwTgM2zUKCyKzre2sUSgg+0K21AYhTIRWW0RRUTae7ne/McfE8eMpLDK+R8FKlgQSpicOQxf1v7+ruPUaimMxtBzeKAzD2pw3PuYUhjGp13Y3L8OwFJOxNBi6rZdg6LreIZdLkkeL4e/niUKhoChLuq2F8fSA1sg9l/V6vcUf153Kz08FwN3l7rutnR0dcXZ0lMoqqjpjZp5/hmFdzRutwDCTPemG2wupmHGKq4D1GFofu8mlkuTRYnh7u6FUKinKkeRRGEehsGLl2UsE//i7UeXTCgqIUCjvw3Ep2X451ejjyiospEdd02xtKyhIB8DLw0MqnBDGJZi3mlmnxTD7PbOSXleLYQ3QTRhmyAtJHi2Hm5sTqcdTJRDCKEFB7Xnx1XSTOy539xD+77UMs4hxVlYCANX8/KTCCWHadmBYH/RlySksh1JCABHhNSjMKKSksESCIYQJyMw8B0D1gAAJhjBHOUAiZe9Trcewv3fafTiWwtJjMWax41uV02MY+5gHpGLcguSiipNfCYCfrwfaYi0ZZzLwivSSgAjxgGVknEGpVFKzenUJhjBHH5f+laUAiLhPx7IUwyLtRn1dlnH7ISBQLqvlkJZHIDysOnqdnrxk2WVGCFOQlZWASqWSbmsh7s05TGQ7O2FepOUR8PE1LAeSeiyVmm1rSkCEeMAyMs7g5e6Ok4ODBEOIuzeMmxf4Fib/4zmPc+evcOLkBRIuJLNnT/zVu6xN5RgleQQiww0JY1ZClgRDiAdMr9eTmXmO+nWCJBhC3BtJHKuI4ycSmDtvI+s2/MM/+06Qk5N/q+XUJHk0JWFhhnFVyYeSTeq4inOLWfLEEjT5mnt+rpCuITR8suG//394xmHWjS17Rym1nRpHH0cCmgZQp38dfGN8paKYoF27viU+fv4dPSY4uDMtW771X104PIN168aWXRfUdjg6+hAQ0JQ6dfrj6xtTyb+6z1NUlEVkaKhcYCGE2Sop0bJoyVa+nTiHnbuOotGUYO1ojX8jf+o1rYZ7iDtutd1IOpDEypdWmtSxS/IIeHu54eHhQuL+RNP6yVis5eTykxWy77ZL9esXWy7KKSLzrGHpL9darqjt1P/epyvRkZecR8rRFM5uOMv2z7YTNSiKLhO7YOtqKxXGhKSnn+b8+c139Bg3t+tb9IqKcsjMPGuoC661UKvt/qsLuhLy8pJJSTnK2bMb2L79M6KiBtGly0RsbV0r5ZySk4+g1+uIjoyUCyyEMDt6PaxctZM3x/7CocOnUVlbEdI9hOih0dRsVxMbRxvDZpSlirOLTe4cJHkElEoF0fVqs3n7QXKTcnH0MY1dKawdrek9tTdazX89DwqF4t8Erji3mJm9ZgJQu1NtWrzW4t+kszjv+srmHlz2Nm99pvWhWrNr9hDWG/6yLmQRNyWO7V9s5+C0g6TEpzBs3TBsXGyk0pT74aCjsDCLsreFvZla7YCVlc1d1mEV/fsvwNm5/KVt7OzKXni7T59pVKvWjBsrQ1bWBeLiprB9+xccPDiNlJR4hg1bh42NS4XHLjn5EAD1IiKkIgkhzEpySgbPvfAN8xZsRGmtovFzTWj2SjNca7hWqfOQ5LFUTHQI6zfsI/1UuskkjyprFWE9w8q8vzCz8N//dvJzolb7Wnf1OgqFAoXimp85pf/pWtOVNu+3IaRrCH91/Isr/1xh9aur6f5bd6kw5cjLS+GXH0IId3e69gdkmS7l5BDdcgJNmjx3l6+owNu7zk2tivdcF0qP3tW1Jm3avE9ISFf++qsjV678w+rVr9K9+28VHrukpDgUCgVR4eFSkYQQZmPjpv0MGzmBhItJhHQJodM3nfAIqZq7aEnyWKpOZE3DF9eBJKq3kLXlruXfyJ9WY1ux5rU1HJhygDbvt8G5mnOZ5XUlOlKOpJB6LBWllRLf+r641nRFoVRYVNyqOTuzY9RIrJTlr4j1wcaN7K4KdcG/Ea1ajWXNmtc4cGAKbdq8j7NztQpOHg/h7+ODh5ubvPmEEGbhr2mrePKZz9Gq9HT9sSsNn2xYpb8T78s6j1qtjlVrdvPUM1/Qpv2LxHZ4iRdf/o4dOw/fajbRA1GvbjAAl3Zfklp+C2E9wlBaKdFpdCRsSbhlGb1ez9E5R/mtwWRW9znF5e9qcf5zP+a32cWUtn9xYdsFCaQ51IWwHiiVVuh0GhIStlTocxcWZpKRcYaYqCgJtBDCLPz86yJGPvkJancbhq0fRqOnG1X5xpRKb3lMTEpnxOhPOHPWnfr1RxMdUxe9XseFS7t5bOgvtGrhwI8TX8bR0e4BJ49BODs7cHHHRcMwL4VU+GtZO1mjVCnRlegozCq8ZeK4edxmzk2BPl1W4eV1/WSH8+c3s2zAGFp+kUXUQEkMqnRdsHZCqVSh05WUjuusOJcv76WkpIDmjRpJoIUQVd70mWt5/qVvcK7hwqClg/AM9zSL86rUlses7Dx69B6LlXoEAwcuITy8N+7uwXh4hFKv3mCGDNnA+QuNGfj4B2g0D3ZfaWtrNY0bhpOZkEn2pWyp8TfIOJVBSbHhGrnVurk78fii45ybomPQwCU3JY4ANWq05rH+K9n8ahzJR5IloFW5LmScoqTEMCHLza1WhT73hQvbACR5FEJUebt2H2X0U59i62nHwMUDzSZxrPTk8dPP/sbGtgcNGz6JQnHzSymVVrRtO46Ei7X4c+qKBx6M5s2i0BZpubznstT6a+h1enZN3AV6cPR1pHqr68eE6kp0bP1oJ507fYe1ddmTjZydA2nddBzbPt0mQa2qdUGvY9euiYAeR0dfqldvVcHJ43ZsbWxoVK+eBFsIUWWlp2fz+NAP0ei09J/bH69IL7M6v0pLHguLivlr+hZatHjjtuUUCgVt247jp18W8aCHP7ZoUReAhG0JUvNLafI1rHl9DfFz40EB7ca3Q22vvq5MxpkMVCnV8PEp/ws/NLQ7V7ZkocnTSHAr/MPqNKmp8bf9Kyq6+25mjSafNWteJz5+LqCgXbvxqNX2FXb8JSWFXLnyD9F16uBgby8XVAhRZb3+1k+cPnOJ2A9jCWwRaHbnV2ljHuPjz2NvXxdb2/LXgXNzq0VKqj3JyRn4+Dy4GZaNG0VgrbayuIkdV/Zf+bdLGgA95Kfmk3ggkSOzjpB+Kh2FUkGrN1sRM+Lm3UWyErLw8ozCmIGi1taO2FGN/LR8XBxc5BOmguh0JUyb9nC55fr0+Zu6dR8ruy5c2f9vl/TVypCfn0pi4gGOHJlFevopFAolrVq9SUzMiAo9h5SUIxQUZNC8YZ8blgsSQoiqY8vWOKb8tZLqLavT7OVmZnmOlZY8pqZm4eho3JZ2CoUSezsfcnLyHmjy6ObqSHR0MAfiTpvUYuGVbfmzy297v3cdb2LHxxLeM/yW+aFSraSkpNDIV9Oj1RWitFLKJ0wFUigUBAa2wMrq9hPPyntPLl/+7O3rgncdYmPHEx7ek4qeVXbmzFr0eh2xLVrIBRVCVEl6vZ6x7/2OTqGny8QuKFTm+UO40pJHf38PMjPPGRlsHbl5F3B3d37gX8BdOjVlz95jnN90njr961hEZa/Tvw5uQdcn7UorJQ4+DgQ0CcCvgd9tkz3PUE8SU5ej02lQKtW3fa3c3CRKHJJw8HKQT5kKrbsqevWacs+LhNep0/+m51AqrXBw8CEgoAl+fg1QKivnY+PUqZXY29nR6qGH5IIKIaqkTZsPsG37ISIHROIb42u251lpyWNYaHXgFLm5ieW2diQlHaRWTdUDTx4BOnd+iHEf/cmplacsJnmMHhZNyCMhd/14R19H3JrA8eOLiYjoe9uye/f+RHj/2ijV0vJoknUhehghIY/c99ctKEjj8uW9tH6oES5OTibZmnD+4kW0Op3Rj/Fwc8PV2VkqlRAW5IefFqBHT6u3W5n1eVZa8mhlpeK5Md2Y+vfb9OjxR5ljmHS6EtaseYUvPx1oEgGJqReMr687Z9aeQVukRWWjkndDeRTQfkI7ZnV8BQ+PMLy9b72O44kTSzmW9gfDXxpsEWFJLyjgvQ0bUBoxfm/z+fM41rHcKnT+/GaKi3PpHBtrkuMd8wsKaNm9O5FGJoMXs7IY9cwzvPLUU/L5IISFSE7OYOXqXQQ2C8SrjpdZn2ulLhL+7JjerNvwLmvXvkFs7DisrGyvu7+oKIdly8bQsb0jPXu0NImA2NnZ0LZNfWbNXk/ykWT8GvjJO8IInuGePDK1DXOGd6FxndeJihqEnZ07er2e3Nwr7NnzI6cyp/Lo3N7YutqafTxsbV1o3elnTui1RpX39gY/v4YWW39OnVppGDYSG2uyx+hub8/qwcb98Plp717y5WNBCIuyYuVOcnMLaDO4rtlP+qvU5NHaWs3sGf/jtTd+YMqUpoSHD8LXtz56vZYLF3Zw+tQsnnqiFW+/+YRJBbpTxybMnLWO44uPS/J4B2rF1mLo9v5s/2I6kxd8jLXGG52uBNwyiBwYzvBnh2DjbGMRsbCysqVOnUelUhihpKSIU6dWEVyzJiFBQRIQIUSVtHb9PyiUCoK7BJv/d1xlv4C9nQ0/fPcyLz5/gcVLtnE0fjNWVioe7VObnj0mEFjN2+SC8kiXZjg42HFs/jFav9sapUrG5xnLKcCJTl934uEv9RSkF6BQKrBzt5PAiDJdvLiTrKwERg14ArWVlQRECFHlaLU6tm4/hHuIO84B5j/W+b59UoeGBPLqywOrRFC8vVxp0zqaFat2kXI4BZ9on3t6vpKiEqysrSxqv2yFUoG9pyz0LMp35MgsFAoY0KOHBEMIUSWlpmWRnJxOza61LWIpOvmZX4ZBAzqwfMVOjs45elfJY9rJNPb+tJeEDZfQ5qhR2msIaOFDw6camvX0fSHu6IdVSQEnTiwlJCiI+lFREhAhRJWUkJBEfn4RXhFeFnG+0h9bhs6dHsLZyZ74+fHoSoxfngM97P15LzPbrsTr1EgGttvBqAGHeLzTHqonv8zi7jvZNG4Tep1egiws3oULO8jOvkCvTp1Qq9USECFElXTlSioAzoGWsTyXtDyWwdPDhbZt6rN42TYSDyTi38jfqMcdnHaQg5+kM/yxbTg4/DeeU622p27dQYSGdmXBjCFst91Oi9fvbScNtYOa/nP7A+BS4862+gvqGPTvY2VSUNVVv/5IatZsAyhwdLy74RVBQR3p33+uoS74Nbivx3/48AyUSiX9unaViymEqLIyMnMBsHOzjDH+kjzexohhj7B46Tb2/b7PqOSxIL2ALe/uYXDv7dcljteysXGmV88pTPquCZH9Im/a2eVOqNQqIvpG3NVj3Wq54VbLTS5yFefrG4Ovb8w9PYebWy3c3Grd92MvKEgnPn4edcPDpctaCFGllZSUGP7DQuY2SLf1bXTs2Bh/P0+Ozj5KUXZRueVPLDlBbd8+uLgE3racra0r9es8zeEZhyXIwmLFx8+joCCDEQMHYiWzrIUQVZiNjTXAnQ1zk+TRPDnY2zJ0cCcKMgo4Ou9oueUv7LhAUFBHo567Vq12XNxxUYIsLJJer2P//sk4OjgwUGZZCyGqOFdXw7aqBekFFnG+8nO/HEOHdObLb2ax/4/9xAyPue1i5sU5xdjbexj5K8UJbZFMmhGWKSkpjkuXdvJY7974eFWN2YlZhYU8u3y5UWUPJyfTo2FDudBCWIga1X1QKBRknMmQ5FFARHgNWraox4ZN+0g+lIxPvbInJbjWciX54GGqVWtW7vOmpZ3EMaDqrYNYmFlIwtYEMs9lYutqS2DzQFxruqJQKqSyCKPt3z8JgFGDBlWJ47W1sWHyxIloNBqjyvcAwmrXlgsthIUI8PfE0dGOtONpkjwKgzFP9WLDxn3s/HonPSf3LLNccOdgNs2cTf36o1Aobj8i4NCh6YSODakyMdDka9g0bhPHpl6kum97PD3rk5SXwq431+MQncPDX3bEM9xTKosoV15eMnFxfxEVFkbLxo2rxDGrVCratWghF08IcUtubk4EBnpzcd8VtMVaVNYqsz5fGfNoTCtCtxbUrh3A4RmHybmSU2a5wOaBKILPcejQ9Ns+35kza0hTbyKsZ5jRiVtuUi55yXloNdr7fv7FucXM7DEL/ZZmPDHsED16/EHz5q/RseNnjB61m/qOHzOr81Iu7pIxnKJ8+/b9TlFRFs+PGiVrOwohzIJCoaBt6/pkX8wm7YT5tz5K8mgEGxs1zz/bl5KiEvZ8v6fsyqNU0OP37myLf5N9+35Dp7u+i0uv13L06FxWbBlFj8ldsbK5fcNv6vFUFo1YxO91pzOr8VamN1zP7zFTWP/Oegoy7t+g3LVvriWgeCAdO36GjY3TDW8YJRERfejZYQZLh68yala6sFwaTT579/5ENT8/BvXsKQERQpiNjh0aAXBs4TFJHoXB8CFdcHd3Zu/Pe2+bIDlXc2bwugGctvmGPyY3ZcOGd9m37zc2bRrHn1PasD/7bQau6ol3Xe/bvl78/HhmtV9BrayXGdn/IKNG7uSJUf/weJc9qLZ1YErLv0k5mlLp5515LpOERXm0avU2t1vAqlq1ZgS5DeDAnweksogyHT48k+zsizw1eDCODg4SECGE2ejQrhHeXq4cnnEYvda8J8TKmEcjubg48OTo7nzy2d/ETYmjyfNNyizr5O/Eo7MfJf10Ouc2xpGXvB0HNzu6tq6HV4RXuZNLLu2+xMYX43i831pcXa9fvNnR0Yc2bd6j2umHmPfoKIZvG4ytq22lnfeZtWcIqdkTtbr8yT316g1mzaJ+uNVyI35+PKnHU1FaKfGN8SVqQBSBzQMtZgFVcTOdTsvOnV/h5OjIE4MHA3DkxAmysrONfx86OVEnLEyCKYQwOY6OdvTp1Yaff1vE2Q1nCeoQJMmjgOee6cu3E+ey/YvtxIyIwdrRuuzCCnAPdsc92P2OXkOv07Nh7EY6xf54U+J4rdq1O5GQMIJd326lzfttKu2c00+m4+1l3C42zs6BXN6RyoHEImKix+PXLAatVsPFiztYN/hHXFrtpesPXbF2spbKZIHi4+eTnHyE50aMwMfTMLnqpffewyYxEXe78rf0yigsJM/Tk/Vz5kgwhRAm6fnn+jJpyjK2fLSFoPZBZttgIsnjHQjw92T0iG5M/HEe+37fR9OXmlb4a2RdyCI/3omgFh3KLduo0RimzZ5E63dbV9pSOWoHNRpNvlFldToN7m4hPNpvznWzzV1daxIZ2Y+NGz9gwZD5PDr3UZRWMmLCkuh0GrZs+RAHe3vefO656+77MDaW+n7l768el5TE/x08KMEUQpisyIia9OzekjnzN3Ji2QlCu4Wa5XnKN/gdGvvWEJwc7dnxxQ6Kc4or/PlT41MJ8HsIY36uODr6YpXnTV5KXuUlzE0COHduk1FlL13ahY9PvVsuU6RUqomNHYfqVDSH/j4kFcnCHDo0g6SkQ4x+7DECfH0lIEIIszVh/FM42Nmy8sWVaPI1ZnmOkjzeIR8fd8Y83YvsS9nsmrirwp9fV6JDqTRu+RKFQoFSoa7Ugbk1WtUguWQzqanxtz9uXQk7dnxNvXqDb3O8Ktq2/YA9P/6DXi+761gKrVbD1q0TcHRw4M1nn5WACCHMWnDtAN58YzAZZzJY8cIKSR6FwSsvDcDV1ZFd3+6q8CVz3Gu7k5QUZ1TZwsIsilRXsPeqvJ1q1A5q2n3SigWLh5CXl1Rm4rh27Ru4uQVRo8btx196eISgS3QhPzVfKpKFiIv7k9TU4zw3YgS+3t4SECGE2Xvj1cdo3SqaA5MOmOUqJJI83gVvbzdefK4fecl5bP1oa8UmjyHuaNzPkpRU/tiuI0dmUbOTHyp15a5kH9YzjEYfBPDn3y04cGAyBQXp6HQaNJoCEhK2MH16Ny5f3ssjj0wsd2cdhUKFtdIdTZ5GKpIFKCzMZNOmcbi7ufHyk09KQIQQFkGttuLvKe/h6+vOsjHLOLXylCSPAl5+aQDVA33Y8+MeUuIrbr1FpZWSVu83Z/nKZykuLns3m9TUY+yMm0Dz15rfl/OtP7I+Azd24ZL/j0xeGMWvU+ryy7RQtqWMpuboEtRqe9Tq8tft02qLKdQlYetiK5XIAmzb9inZ2Rd598UX8fLwkIAIISxGtWpeLJr3MQ7WNswdMJeTy0+azbnJbOu75OzswPhxoxk64iPWvbmOAQsGVNiM5/Ae4SQfSmb6pEfo/PBEfHyiUSgMz63TlXDy5HLWbnmRTr+3wrWm6307Z49QD7r+2BWA4rxirGysUFopKSks4dAfU8nNTcTR8faTIS5f3otDaAm2bpI8mru0tBPs3j2ROqGhPDVkyB0//kRaGi+uXPnv/+cUFRGfk0OXwWWPq/3k7beJjoyU4AshTELjRuHMn/MRfQe8w+x+s3nk+0eIGR5TaSukSPJYBQwa0IFff1/CtqWHOLPmDLU71a6YJ1ZA67Gt8al3nGXje6FMCcDHK4aSkkISU/fgWK+Ifss741PP54Gdu7XDf2s1WtlaEf1kOBtnvUfXrj+X2XWt0RSwbsNbtPyhqUlcP71eR1FR9h09Rq12QKWqmvsx6/V6ioqy7sv56vV61q8fi0aTzydjx2Jne+c/FrKKiigsKeHXbt2MKv/8ypWkZWTIB5MQwqS0i23AqmVf0rf/Oyx5YgkXd1yk4+cdK3WDD0keTTl4Vio+nfA0bdo/z+pXVzO61WjU9hWUWCggrEcYIV1DyDiVQfrpdFRqFW0jOuBczdnkYtHk+SbM3TKPdeveom3bD7Cyun7R54KCNJYsfRL/3pqKS7LvUWbmWaZOeohqTsbFMzk/j/aP/E5oaPcqWV8LClL59ccIAp2NO9/Ugnzadv6FsLA734P69OnVHDu2gC6xsTzSrt1dH7ODWk2Ikd3dTtay+LwQwjQ1aRzB1k0/MPqpz1j7+17ObjhLh086ENEnokq2QkryeI+aNa3D4Mce5s+pK9j+xXbavFexu70oVUo8wjzwCDPt8WJWtlb0ndGHtW+u5Y8pDxEZOhA/vwbodBoSErZy/Pwc6r8QTNOXO5jMG0Wv11HHzZ6tI4caVX7MsmWka6vuRB+9Xk+kmx3bRxl3vm+uXcuZuzjfoqIcVq58EXtbGz575x2UShlaLYQQNar7snThp3w7cQ4TPv2LuQPmEtgskOavNyekSwhKddX5rLTI5DEzM5fMzBw8PV1xdLS7p+dSKBR8+vEYVq/Zw7ZPtxHeK/yBdic/SGoHNV0mdiH9xXROLN1M/NF5qNQqfNv5MqxbHxx9HeXTwwJs3vwhaWnHGffaa0btQ/3amjW43qJbO6OwEDsr+X0rhDAfNjZqXn/1MR7tG8u4j/5kxqy1zOw1E7dabtR9rC6h3ULxifbByta0P/ss5pNZq9Uxa856vv9xPgkX9FhZOaHRpBFVx5VXXhrAwx2b3PVze3u58vmnYxg8bDzLn1vOsA3DUKost7XFPdi9UrZuFKbv0qXd7Nr1HTF16vDq00+XW/7nTz8lJzf3lvcdOX6cGd99J0EVQpidWrX8mPz7W4x9ayg//ryAWbPXs3n8ZrZ8tAVbd1t8o33xCPXALciNrIQskzt+i0geCwqKGDJ8PEfi3YiNnUGHjhEolSp0uhIuXdrFE2Nep1+fPXz+yTMo77JLdWD/9sycvZ4lS7ex98e9NHm+ibw7hEXRaotZvvxZVEodP0yYYNQkmdo1atzy9hpNmpCfnc1DspWhEMKMBdcO4KvPn+Pj8U+ycdMB1m/cx4aN+zi66zxn15812eM2++YxvV7P8y99w5WkxvTvPxdv7yiUSsOi2kqlFYGBLRgyZD3LVxbz1Tez7j6QSiUTv34RZ2cH1r+znrQTafKuEBZl69ZPuHx5L2OGDaN5o0b39Fx2CgVLBwyQoAohLIKNjTWdHm7CpxOeZte2X8lMWc7Zk7PZvOF73nl72NVihZI83id79h5jzdo02refgEJx651YrKxs6NVrMl99u4LEpPS7fq0aNXz56vPnKM4pZuGwhehKdPKOEBbhwoVtbNnyEZGhoYx77bWK+XBSKCSwQgiLo1AYVnOpWcOXVi3q0aRxxNW7SiR5vE9mzlpHg4ZPo1Sqy8n6XahdeyCLl9zbdoMjhz9Cn16tubjzIhvf2yjvAmH2CgszWbhwONZqJZO//hpnR5kYJYQQ5szsk8e4Q2cIDDRuC7/q1Vuy/8C9bR+kUCj46YdXCKzmzbbPt3F+03mpZcKM6Vm58iXS00/x7v/9H01iYiQkQghh5sx+woxGoyu31fEqlUqNTqe/59f08nRl0m9v0rnbaywcsZAn9jyBvYe91DZhdg4dmkFc3FTat2xp1OzqO5VbXMzRFOP2js8pKpILIoQQkjzeu+DaviQlHcTVtWa5ZRMTD9CxfUCFvG6H9o14+aX+fP7lDBYNW8SAhQNQWsliycJ8JCUdYtmyMXh5uPPHl1+iruA1GV1tbXG2seHNtWtvWy4hO5siR0dCgoLwdHeXCyOEEJI83pu+vdvw6puTCAvrDpQ9AF+rLSY+fgbfffVKhb32+P+NZt/+E6xb9g+bxm0idlys1DgTcyknhy+2bzeqbFxiIoFBVft8L9/B+e6+dAlP/1vfV1CQzty5A9Bq85n67Z/UqFatwo81xN2dxYMGlVtu8v79XI6IYOwLL0iFFkIISR7vXaeHm/DlN7M5eHAa9eoNKbPc5s3j6dIpgPCw6hX22tbWaqZOHkvTlk+z9ZOtBDQOILR7qNQ6E+Hg4E2dFhPYaGR59+je+PpW3TF9NjZORLX82Ojzta8Dfn4NbrpdpythyZInSE2N56M33qBzrPwoEkIISR7N6QStVEyd/Dbdeo4lK+sCTZu+hFr93/jDgoJ0Nmx4FxenPXz1+fgKf31/P0+m//U+D3d5mUUjFjFi2wg8wzyl5plEMuVCo0ZjLOfNbmVXIee7bdunxMcvoHeXLrzx7LNSkYQQwsJYxCC8agFerFv1OVGR+5k8uS5z5w5k+fJnmTGjB9OnN6Bnt2KWLvr4nve5LkvL5nX58rNnyU/LZ3af2eSn5EvNE1XSkSOz2LTpf0SEBPP7F1+gUqkkKEIIYWEsZm9rDw8Xvv3qBcaPy+fAgZMkJ2cQGBhNdL2XsLFRV/rrP/1kT47Gn+P7H+czp/8cHl/+OFZ2VlIDRZVx4cJ2Fi9+Ag9XZ+b++ivurq6V9lqFJSX8tHevUWV3XLhAWESEXCAhRIU7ePA0D7V46oEeQ1ZWniSPD5qToz2tWkZX2AWdMm0lq9fs4cqVNJyd7WnZvC4jhj9CUK3rZxooFAq+/Ow5EhKSWLx0G0ueXELPP3uiVMkMbGH60tNPMnt2P1RKDTN/nkRkaOWO3X129GiS8o1roQ8KC6P1Qw/JRRJCVCQ9cCEnN79g9554UzmmXEkeq7h1G/5h1JPfUK3aQCIjv6FOVBD5+ans3b+KX/94lzdf68RLLzyK4pot1qytrfjrz3do9/BL/DPtIK41XQ0zsGUXNmHC8vNTmDWrN4UFyUz6+mtimzev9Nd8bcwYCbwQ4kEqAhpLGCR5rDDbdxxmyPCf6d17GV5edf693c7OAw+PMKKjh/LDzwPR6Wbzyv8NuO6xzs4OLJj7EW3aPc+WCVtw8HagyfNNJKjCND89i7KZObMnKSlHGffaawzp21eCIoQQFk76TO+QRlPCU898TdeuU65LHK9la+tKv36z+Pyr1Zw5c/mm+wOrebNo/sd4uruw8qWV7P9jvwRWmJzi4lxmz+7HxYs7eWbYMN56/nkJihBCCEke79SmzQfQ6uoREHD7MVa2ti40bPgif0xedsv760YFsWThJ7i5OrH06aUcnnFYgitMRklJIQsWDOHMmTWMGDCAb/73P1RK+bgQQgghyeMd27L1IMHBnY0qGxzcmc1bD5R5/0NNIlkw5yMc7e1YNGIRxxcdlwCLB06n07Bo0QiOHVvIwF69+PHjj7GykhEuQgghJHm8K5lZuTg4eBlV1tHRl9S07NuWad0qmrkzx6FWqpj3+DyOzj0qQRYPTElJIYsWjeLw4Zn0ePhh/vjiC2ysrSUwQggh/iXNCXfIz9eD4yfPGlU2Pf001at5l1uuY4fGzJv1IX0HvMuCIQvQ5GuIHhotwRb3lUaTz4IFQ4iPn0+3Dh2Y/sMP2NvZSWBuo9fIkZw+fvy6VRXKTMx1OmLbtuWHCRMkcEIISR4tycMdG/PTr1No2vRFyltj5+jROfToatxM/y6dm7J04af0fnQsS0YvoaSwhIZPNpSAmym9Xk9KylF0Oo3Rj3F09MXR0bdSjqe4OIfZs/tx+vRq+nXtytTvvsPO1lYuVDnS0tNZ1KMHPo6O5Zbde/kyv2ZmStCEEJI8Wpr6MSFEhBdz+PBMoqIGlVkuPf0UZ85MZeTwn4x+7naxDVi66FN6932b5c8sR5OvoemLTWUdSDOk1RYzZ0YXYtzVqIxotbqUk4NXnedo3frdCj+WgoJ0Zs7sSULCVob268dvX3yBtVotF8lIdmo1DkbEy1bGjQohJHm0TAqFgl9+fJX2D78BKImK6s+N2V1SUhwLFgxk8m/P4ebmdEfP36pFPZYt/oyefd5i9curybmcQ/sJ7VFayfBUc+NmZ8eSQY8alVRMiYtjSlbFvr5OV8KVK/+wdOlTJCbG8fTQoXw/frzsVy2EEEKSx4pWo7oP61Z/ypNjPmfKlJ+IiOiLu3sIeXlJnDq1ipKSXcyc9iJt29S/q+d/qEkk69d8S6++b7P98+1knc+ixx89sHaUiQui4pw7t4Fp0zqh1+vx9/UltlkzMrOzcXd1NWoMnxBCCEkexR0mkCuXfs7ef46zes1uEi6sIdDfmSdG1KdD+9FYW99bt19kRE02r/+evv3fYfvsw2RdyGLggoE4+DhI8MU9O3ZsEQsXDsVarcbd1ZXE5GQGjBmDSqkkpk4d2rdqRfuWLWnaoAFOjo6STAohhJDksSIoFAoaNwqncaPwSnl+X193Vq/4iuGjJjB3/kYmt5rMo3MfxaeejwRf3BW9Xsvu3d+zZs1ruDo78vf3k+nYujVHTpxg4/btrNu6la27d/P5Tz/x2Y8/4mBvT+Po6H+TyZg6dWQijRBCSPIoTJmDgy3T/3qPGtV9+Pq7OUxqMYluv3Sj7qC6MpFG3JGiohyWL3+GgwenEVyzJnN/+43oyEgA6oaHUzc8nOdHjqSwqIj9hw+zfts21m3dyj8HD7Jxxw7e/ewzPN3dad6oEe1atKB9q1aE1Kol60AKIYQkj8LUqNVWfPHZs9SNCuL5l75lwZAFXN57mQ6fdEBlLZMbRPnS0k4wd+5AEhP30yU2lj+/+QZvT89blrW1saFZw4Y0a9iQsS+8QGZ2NnsOHGDtli2s37aNNZs3s3j1ahQKBYH+/rR66CHat2xJm2bNqBEQIBNuhBBCkkdhKoYN7UJ0dDD9B73Pzq93khSXRK+pvXAOcDbL89Xr9BRmFKLT6rBxtsHKVqrr3Th+fBGLF4+muDiTt557jv+9+irqO1iKx9XZmY6tW9OxdWsAriQns33vXtZv3cqGbduYvWQJf8+fj5VKRUhQEG2bNaN9y5Y0a9To/9s77/Aoi+2Pf3Y3vZGQAiQkQIAAhprQQw1F6gVBEFEuTcWLiFz9ieV67V6uioIFUVFREC4iTUWUXgKh9y4lhJoQEkIgPdn9/TETWGJINhAh5XyeJw/svrPvO++Z2Z3ve+acGar5+Um8pCAIgohH4V7StEldoqOmM+qxSfzyazQzms+gz+d9qPe3euVmGjs7PZtdM3axf/YhzAnu2BldyDDEUy3Ck3YvtMM31Fc6gg1kZV1lzZp/s337NDzcXZn90XQG9Op1x+et5ufHwF69GNirF7m5ucSeO8eGLVtYHRVF1LZtfDF7NtNnzcLZyYn7QkKIjIiga/v2tGjaFK9KlcrXA450M0EQRDwKZQEf70osnP82kz+cx5vvfMv8gfMJfzKcLpO64OjuWKbv7VrcNX58cCF+mT0Y0OYTvLxqYTCYyMxM4dixZfzY/VUi/hNKk+GyfWNhxMXt4aefRhIXt4eWzZrx3dSp1K9Tp8SvYzKZCA4KIjgoiBGDB5OZlcWxmBjW6njJzTt28P706bw/fToe7u6ENWxI1/bt6RwRQdPQ0HKx/eGSI0fwtCGJ6ERSEpQz8SwIgohHoQxhb2/HSy88SudOzRj52CS2T9tO7PpY+n/bn2rh1crkPZmzzfw4aCFNvV+hWbNRWLtSHR09aNhwCDVrdmTOy/fjEehBrcha0hHy29Ccw/bt01iz5hUs5gxeHDeOV//5z7uWIe3o4EDDevVoWK8eT48aRVp6OvsOH2bNxo2s2riR7Xv2sG7zZgwGAz5eXrQOD6dL+/Z0btuW0JCQMhcvOWrIEM6cP89FWwrXrMmA+vWlkwqCIOJRuLe0bhXK1k1f8NzEaXw981e+afcNbZ9vS/uX2mPnXLaad+/svXhd7UCzLjcLR2vc3KrRr+93LJnQmzF7HsdgLLtz9VcyMhj98882bU944vJlHOuEFVomMfEPli17ipMnV1OzegDffPghnSMi7uk9ujg70zosjNZhYbw8fjwpV6+ydfduVkVFsTY6mt/XreOXlSsxGY1Uq1KFjm3aEBkRQee2bakVFFQmxKMgCIKIR6HM4eHhypfTn6dH91Y8PWEqG97awJFFR+jzRR8CIwLLzH3s/34/3VsvoqjgzapVm+GaGsqFXRfwb+5fJtvMZLKnV79Z5ORk2hQ3FwxUrlzwtLPZnM3mzR+yfv1b5Oam8+jAAUx5/XV8KlcufX3V3f168o3FYiHx8mXW63jJNZs2MXfxYuYsWnR9OrxL+/Z01ZncpfF+BEEQRDwKZRaDwcDAAR2J7BzGcxOn8d3s35jZYSYtxrYg8u1IHCuV/ljIpENpVO3Q1KayNWp0IG7P9jIrHg0GIzVqdLzj81y4sJNffnmCCxd2ERwUxCdvv02vLl3KTJ/1qVz5evKNxWLhzPnz19eXXL1xI1/Mns3ns2ZhMpnUzjft2tGlXTvaNm+Om6vstiQIgiDiUbhjvLzc+WbGiwx9uBtjx33Atk+3cXjRYSLfiaTJsCYYTKV3mteQa3t3tLNzwpxjrrDtnJp6kfXr32Dnzi8xmQyMHz2atyZOxMPNrUw/AAUFBDBi8GBGDB6M2WLh8B9/XBeS1jvfODk6qnhJncndtGFDnBwd5QdAEARBxKNwu3SNDGfXtq95Z9IsPvp0AT+N/Ikd03dw/4f3E9g2sFQu6+NUPZeUlLN4eFQvsuyFC7tpGuxV4do1JyeDnTu/YP36t0hPT6RF06Z8/OabtA4PL3f3ajQYCK1Xj9B69Rg/ejSZWVns2r//umdyx549rIuO5t/vv4+3lxdtmzena/v2RLZrR0hwMA729vJDIAiCIOJRKA5ubs5MemcMo0b25oWXpvPz0k3M7DiThkMa0vnNzniVMvFVv3999qz/lg4dXim0XGpqPOeS1tKvw+gK05Zmcy7Hj//GqlUvkpBwkICqVXnrnQ8ZNnAgdnYV42vs6OBw0843KVevsmX3btZERbFq40ZWRUXxy8qVf9r5plPbtgT6+2MnO98IgiCIeBRso26d6iyc/zarVu/ghZc/Z/ec/RxZcoSmw5vSdmJbPGt4lop6thzfkq9nfkbdur2oVi3sFiIqh2XLnqL1i80qxI4zFouZmJi1bNjwBqdPb8TV2ZkXnnqKiWPHUtnTs0L3aw93d7p36EB3vfNN3MWLbNqxg1VRUazfvPn6zjcmk4mQ4GA6tmlDl4gI2rVsSRVfX9n5RhAEQcSjUBgGg4FuXVvQoX1Tvpv9O+9NnsP2z7azd/Zewh4Lo82zbfCofm+3OXT2cqbfnJ4sfPgBIltPpn79BzAab3TR5ORTrFjxLM4RsbQY26eci0YLsbEbiIp6m5iYNTjY2zFs4EBeHj+eerVrS4cugKpWO9+YzWZOnTmjdr7ZuJGobdv48vvv+XzWLJwcHWlYrx6R7drRpX17mjduXOGFuCAIQrE0RSHHegLLpkx+mgnjB4mlyhmpqRl89c1SPpgyjzNnL+Lg7kDY6DBajmuJV+17O5196cglVr2wmpTdDvj7tcLRsRIXLx7gqt1BWj7blGajmhWY+JNyNoWpNaYSHjaG3r2nl8l2MZtzOHVqHZs3f8CJEyuxszMyqE8fXhw3jkaywPRtk5WVxR8nT7I2OprVGzcSvWMHCYmJgPJghjVqRJd27Yi8Bzvf3NepE4ePHdsNhElLCYIg4lEo9Vy7ls4XM37ig6k/cOFCInZOdtTvX5/WE1oT0DLgniXWWCwWrp67Svy+eLLTsvGq5YVvqG+hU9VlWTzm5KRz6NBCtmyZQlzcHuxMRvp2787LTz9NWOPGyARrCT88paWx/8gRVkdFsXrTJnbs3cvVa9euLx/UOiyMru3b06ltWxrUqYP9X5h8I+JREISyhkxbV3Dc3Jx57p9DePKJfnw/dwUff7qQA/MOcOjHQwR1CKLV+FbU6VkHO8e721UMBgMe1T3u+VT6Xy/e49i7dxY7dkwnOfkUri4uPPbwEJ557DHuCwmRuLy/CFcXl+s73/zrmWdITklh2+7drNm0iTWbNrE8b+cbkwn/KlVo36oVXdu3p2Pr1tQKCpJ2qXhUBt4H3gOOijkK//kG/gWcB74Rc4h4FMrzYOrqzJjH+zFqRG+W/b6FD6f+QNS6fZxaewqP6h40+XsTmoxognddbzHWHWKx5BITs5adO7/k+PFlZGWl4ufjwyvPPMM/hg/Hv0oVMdJdxtPDg+4dO9K9Y0fMFguXkpKI2rqV1Xobxf8tWcLcxYuxs7NTO9/oKe4Obdrg511mvxNDUd7O/7vL1w0HfIANQLrV+x7AG0Bz4AfgU2AcULOE6hiszzMd2F+MzzkDvwDr74Fw7Ak8Afjqa5fGJSa6AAOBF4CrgAWYCWwDrgAL5ReG1oC77kNZNpRvAAQBO4DE0vqEUFinlWnrCitwLOzYeZSvZy5l7rxVXL2ahtHOSI0ONWg6qin1+tbD0aN0Lcpc2qetr1yJZd++OezZM5OkpBMYDNAmPJzRDz/M4L59ZceUUorZbObshQus2bSJVVFRrIuO5lxcHAAO9vY0atDg+s43ES1a4OriUqzz38Np68+BwSiv2t1kmR5fgoEYq7Fovh405wDJWuT9BrQESkKhtwU2AX/TYtBW3gX66PbJvMvifoYWYqeB6sD4UvgV+SfwIVAFuJiv/p8ATYCzFfxnZBsQCtTKZ6Nb8THwNNAdWFkab0g8j0LBTxUGAy2a16dF8/q8/9+xzF+wli9m/MTOtUeJWRODvbM9IX1DaDCwAXV71cXBzUGM9mcJzpUrZzh0aAGHDi3g3LmtWCxmfCpX5pnRo3hs6FAaShJMqcdoNN68843ZzOHjx1W85MaNbNiyhV3791/f+aZNePh1MRnWqBEODvLdyMf3wBbgstV7QUA/oC+w3Or9wfd4nKqlB/GRd1k42qG8sFOAV8poO89DeSP/DYyp4H3+S8APSLV6zx6opB+Ucgp4wEoETpTWGxLxKBSJu7sLo0f2ZvTI3hw4eJK581bxw/w1HPrxEAfnH8TB1YGQviHU7VOX4MhgXKu6VtiYMLM5h8TEoxw/vpyjR3/m9OkoLBYzbq6u9O9xP0P69aNvt244OzlJxyrDYjI0JITQkBDGjx5NVnY2u/btY/WmTayOimLbnj2sjY7mlffew9PDgw6tW18XkyG1a2Nv26LuHYADQFI5NOHcW4g0+wK8MlfvcV2fBFKAJXf5uj5AQGkWD7b8HAJfA+8ArwLxFfhn46sC3msFRAFNgb35jv2u/0otIh6FYtEwNJj/vPUE77z5OFu2HWLR4vUsWrKBgz8c5MC8AxhNRqqGVaVOjzrU7VmXKk2rYO9cfreJs1gsZGRcJjZ2AydOLOfEiZVcvqx+711dXOjbrStD+vWjV2QklTw8pAOVQxzs7WkdHk5rvfPN1WvX2Lxzp9pGMSqKFevX8/OKFQAE+vvTQSffdGrblur+/rc67SnUdNVA/f870rtWg3lxMAG5d1CmONe1v4N7+6s2uTcBg/QgnnmX62Gnr38v778k+AXlPe1N8ZNnDPrPXEQZow39tCTI84hY/oLv2r2qr4hH4S73SoOBNq1CadMqlP++8yQ7dx3l9+Vb+X3FNvbuPcaG7efZ8NYGnL2dqRZWjcC2gQS1C6JaWDWcvZwpq2vPmM05XL16gXPntnLmzCbOnInm4sUDZGenYTKZCAoIYFCvR7i/Uyc6tmmDt5eXdJYKhrub2/XkG4D4hAQ2btt2fU/u+UuXMmfxYkxGIyG1a3Pm/HkAV6AqyvNmRsW4ndaeiX7Artuoyn3Ac6iYQQPKk/klsKaQz4QCjwNtUFNqu4C3gMNWZZxRU7n9UIkcCUA0Ku7tAtAQmAg004PpCeBHVCyjGegGRACv6/MNB7rq/z8JxGlPzCLgQdSU9of5RN1w/ecNbEZN8Z7NN7aNQMU3ZqCmAc8U03519LWjCynTAhXz1xTlJV6AitXME5tVURna7+njT6BiTAuLXbwf6KHvsz8qYegoNzy2jsBjwEMoD+V+YKq2Qx6uug3W52vvrkB74L/cSFYahkpSegX4B9AJ5W39tgDvl7u+3za6neYXIXLPart3sVE8GnT9Ruj+Yw+s033Q2nPpq++vs67TOWAV8BE3Tw0XxFDdBj8Afwfa6X76DfBTAeWDUIlWHbVIXaXb81K+fvAc0EiLtT90f89LFuoH1EfFzzprG9bRx/6h782sj2egksr+pu8nCZUoZQ98UYAYHKHb4Gv9ujLwIipeMkv3ySlAtohHoVRhMhlp2aIBLVs04NVXRhB/8TLr1+/m9xXbiNq4l9h1pzm58qTqcM52+NT3oVpYNao0qoJvqC9+Df1w8XbBaG8sVfeVnZ3OtWtxJCQcJD5+L/Hx+4mL283lyycxm3MwGAz4eXsTGdGSru3a0b1jR+4LCakwe00LtlHF15eBvXszsHdvLBYLJ0+fJmrLFlZt3MjGbdtIS08HCEF5GPcBa/WA/4MWD6uBR4Ffi3HZzsBi4LgePHK1QPhaD1q38tRM1cJnNipTdqwevFsCsbrM+1q4vKsH7Vq6nov1ALUW2K3LmfRnH0XFOuYJmOetxGN11BRtnmBztxIKA1CJLh9aeXY+0OLpcy2wB6G8tG1RcZQGPVg+BezUtuyOyngtDg11/W+Vmf0wKqFlFfCZFjQv6UG/txZn7lqcHdaCO4Ois499tE0NQKAWQ3nT90768y31g8A5IFL3kbFa8AG4oKaK384nHrtoYfGRlXhsATyiBfAx4IgW7f21yN9pde1lWjiu1OJxgn7guOXPqD5fExtt7qhF3GZgmrbBi7oeEUCa7gM/amE+VdvmPmCItklR4vEhVFjIOH1vKVrc9dJ2mG9VtqnVg8d3+tojtX26oBK+gnUfWANM0vfQVtcnr6376mu8q89RW//liVMPLR7zBsFmuv1ma/HoB7wGLM33kOSDSq55V7/21n3BVdvCQbdRTd0/RDwKpXiw9PNi8KBIBg+KxGKxcPZsAlu3HyI6+gCbNu/n6NHT7N4dd7280c6Ie4A7XsFe+NTzwaO6B17BXlSqUYlKgZVwcHfAqZJTiXsrzeYcsrPTyMi4THJyLMnJp0hOjiE5OZakpOMkJf1BamoCFouagXB0dCTI358eHf9G2+bNiWjenJDatSV+UbAZg8FA7Ro1qF2jBiMeeois7GwaRUbyx8mT8cBWLXAm6r9r+mOeWgA+rcVfUdNQ7qgM3R160LJeDieQwqf4RunBKe8aq7RoHItKfrBDZUrPBSZbfe5dPdD3Bry0Z+2IPjZTD6i3qvc7qGzUzvr/a4oQxWP1IP+jfm8GatmfUVpYRugyM7UnM8/j8ly+OhdFLf3v+Vt4oz7TnqDnuTFNOU+348uoRJE8XtWD/Ov52qMg5miPYU8tuK29rs9p4ROhHzTQns7/6PNvAE7eRtesrL2Mk6zuI1r3n51WXuEI3Q+nWXmBv9dCqTDvY3MtPjOKqEeW9nxaC6QYYAXwgLaNp7bBY/m8mW8Uw7tmp+u8R78ORXn5J6HiW7O08PpGe317a+Ga19+idNvfr+vroPvcBV3ma93nCyJVexL76vt4iT/HPOZnru5PQ7XXM4+B2iM518oGPvp3JM+Gv+i2nKo9oiIehbIxWAYG+hEY6MeDAzoBkJx8jdjTcew/cJLdu//gwMEY/jh+lnMbz3Jq7ambPm80GXFwc8CxkiN2TnY4V3bGs6Yn9q722DnaYe9qj72zPS4+LuRk5IAFEhIOs2vXV2RlXcVsziEj4wpmczYZGcmkpV0iNTWBjIwk0tMvk52dTlZWCmbzjfHUzmTCy9OThvVq0rh+Vxo1aECjBg0ICQ7G19sbO5NJGlYoERzs7TGp/nReeyZcUVNfXfQA46pFmZP2tAUCbxYhAHtocTOqAKFS1NRt/uMXtBfwPv06V7/uqz2h27W3L9dKaBmthFqMHohLKlN5BGoafKv2WOaxRNvsAy0KMvWgbC0mjhTzWj7634LW2XtUC4YPuDm+7bAWXo+jplrz2I5aODvnDu7dhJriXGwlHNGifLIWdcO1h6q45AD/s3p9XIsl6yWShgKHtFjFqj8cK+LcKbr/utkgHs38eVmfQ7oP5fXBNFSG8ihUKMYh/aCVVYz7TQQOWr0+qMXV69pLuh01nd1MeyTTrMpeQXm2Z2jReV4Lxbw1RGN1vyvJ7PxYbsxATLbyUg7XD1sx2r6DrNox7/uRpEVyBxGPQpnG09MNT886NGlch0eHdgcgOyeX5MtXuRCXyMmY88TGxhNz6gInT57jQlwS6RmZJCQkkxAbz7mt57BYbu18iY1dT2zs+gKPubq44O7qiqe7OzUDquBTuTJB/v7UqF6d4KAgagQGEuTvj4e7e7HX6xOEEiBVe+FaANW4Oej9PGqtwSYUHgPZVA/Sh26zDo5awDbQdXDjxnSaRQvDJdpTdVVf51393g5urFE3AhUPuUILp3MlYJ/6qGnC3fneN+qB0aDrfRbb1tMrDCcrgZSflqiYtwsFHDushU2g1Xs/3aFwBDW1H0DB0+iJWtQ3u81z51B0Zn8D3ca3mwxSnLkjd92P66Kmd01WfTBD968vUdPbyVrsvaq/O7fLOl3HOvp8LfT7Bdn7kK5TI5QH/DtUHOPTuh2WojyFl0rwt+FrVBxwc26sG9lKC0a00PfVD53DCtB6Jbr7hIhHoVRgb2fC19cTX19PGjeqXWCZ9PRMMrOyMeeaiYtPIjEphazMbDIyszCbzcTHX2bM2Pfp1qEj40ePxsXJCWcnJ0wmEx7u7ni4ueHi7Iy9vT0ODg6YjEYxvFDacEMtrNxXe9e26YFshxaPtngzLNzIVC3u4P64FnoW1FTaOf6cELEZFQ/YRw9ePVDT6r20UHwWNb3YDRUf9xDQGBUHlnGH9snWA2hB6wbm6npbKJls17w4Qwf+7MG16PHTjj9PlRa0sOflEqiPJZ+ozd92d2NBUfMd1N2Wzzpq0fWEFsQHdDvk78tzUFOxfbSQ74laH7QFymt6O+TPlC7M3o75hPdjqOV4IlFT+6NRntLOlFzm9XL9GzBc/y4MQ8Wd/m5Vb7O2XUHx0SW6TqmIR6HM4OzsiLOz+s5WrvznZW/Onk0ADNQKCqJXZKQYTCiL1EZNS42l6Ni4W7FdD3hhFC/Jpicqlu0JPTjnTQP2KaBsAiqmcKb2Eu1CeRpX6AFsu/5De0bm6/pE36F9dmmRmlWIED2shapPPs9PccX0JSuPzpUCbNwH5QnN75lqr+1zGpWoUFKc12K+QwHH6qE8SztteEC4XQ6jPJsGbo5fLeqc3rqtbFmzcwIqXvZvqLjCXJT3e2gBZWP0gxYob+EeVCLLf2/z/rrp6+VNZ+/Q/3YoQJC216Jxr9WDyyb9Z0DFh36G8pyW1JaW6ajYxtGo+OOh+nXelPpFVNhJBDeHIPwliOtFEASh9LBXD17pd3CO5aj4vne4EbcHyoP490KcBrX0IGi9/64PKkHBmh7c7OXK0F6NbKAGystoTZ7wKomlQj5DZdm+km/8crcSuXO1eH7TqowRlXBRHPK2TqxRwLHv9H1NRXmLsRLKPVH7cmeXcN/IRWVJR+p2zMMFFYOXxo0EEotVm+fhpD97u8xBeZxHWb3nhkoaKYxALWpsiUkMRk1Db+NGuEDNfH3WpG1s3f5XdXlb4x59teDOoxVqynmLlXiMQnn/X8v3ENAI5V1fjpq+rpfvXBZ9DxRRH+vvGKj45qLWO/0WtYTWi/p7MMvqWKbudyO1uLWmPrcf0lAg4nkUBEEoX2SgpoqX6cFttfaGtEJlQv9sNbhZs0IPwOv0Z/MGaesBzZkbS4CsR8XJtdKenzEoj9/3KA/YXi1Y+uiBeG8J3NshlGf0Sy2EdurBtzvKy7pUC4DpKO9tGMpL2FSLrOIK+WwthtfmO3YOtVTPDyiP3AaU568Tysv637+obaegYg+/0SLhHCqxwxOVgZ6X8HQZFff5d1QySS4qFMLzDq79BWq7yC9Ry/jEotY+LCz730ELq2U2XuNnVOjEVt0PvVAeQWvvZoAuF4vyZGegln9K50YGflGYUaEWG/SDR38tQMdY3U82yru3TPeFFVozdedGXCu6flO04D2AWnanr+6Lpwqpw0Fd989QXs7OKM9pdBH9fzPwjLZRfq/3x6iY6JXaRnGo6fN2+l52l1RHLCxttC7wSI/urWjdKlR+joVST0pKGlM//pHwxo3p07WrGEQoE0z79lsuJSXF6cG5pIjXXokMVNaldaxgXvZwjh6MtujXSXrg8kXFdJ1AZSzv1wPjQW5k5V5ErT3npwfMvMHsoB647AF/PcbMQ2Wi5k2v5aKm8jbm89bE68Hceoo4R19/u9V7+1AZx5VRXsFEVNbz+1YD/3I9cFfRHp5pqJi0OCvRWxRX9GDuzs1r/+VxEuWNQ9/redRSKZPyCaor+r6Ksz2fRYuZDdycaGTWomCbtr2HtvcT3Jhmzfv8Bl0mUNfhPS2wT2oxn2Nl48OoKVfr5KBU3UYnrMr9oEWan7b7v7SwPqbLZhfg8XpBi2lbEriO63P5oTyLu1EJWud0nU/pe5mn+3ZVLTDXodb1jLXhGg9ru0Xqhx5/3V9G8eftIC+jvH2XtWjN0A8mE1BZ5Gi7R+kHpQDdRjNRyzVlW9nuoP6OXB+ytMD3QoU5TNV906zbIW+TgPzhGUf1Q8JX/Hl1BDMqqWkvyltaRf9/grZRiVFYrEJPYNmUyU8zYfwg+YUXSj1nzyZQo84gnnj0UaZPmiQGEcoE93XqxOFjx3ZrL5lQungBNUUeYCUWBNt5FTUdXIc/x43eK35CeefqUvKhBRUGiXkUBEEQhIL5BuW5HCamKDZOKG/el6VIOAoiHgVBEAThLyUBtXTMS6jpRcF2nkZNv74vpih/SMKMIAiCINyaL1AJB5+g1tjLFZMUSTNUnN0ACk7Oupd8goqXlXa8A8TzKAiCIAi3Jhc1/XoEtauHUDgGLbYf4uYEkdLCKlQClFma6vYRz6MgCIIgFE4G8LaYwSYs3Fi8WyiniOdREARBEARBsBnxPBb3kcpiYdXqnfz622ZiTp3H3c2FVi3vY9DATlSt6i0GEgRBEAShXCOex2Jw+nQ87TqNY9yENZw9/wD+AdOwd3yFhUsqEd76OT77fDEWi0UMJQiCIAhCuUU8jzZy7vwlIrs/T3jzD+jWfcBNx2rU6EDLlk/zwdShpKSk8eLER8RggiAIgiCUS8TzaCMTnv2Y0Ib/pkGDAQUed3HxYdCg+Xw8bRP7958QgwmCIAiCIOKxonLi5Dk2b71CkybDCy3n5ORJmzYvMW36YjGaIAiCIAjlkns2bX3lSirr1u/mwMEYMEDjRrXp2KEpHu4upc5IURv3UatWD4xGU5Fla9fuzq9LX8dsNmM0ijYXBEEQBEHE4x2Rm2vms88XM3nKUnx8OhMQ0BKLxcyCRVu4kjyDl14YwOiRvUuV8Dp3LgE3t8Y2lXVx8SEpKZXcXBGPgiAIgiCIeLwjzGYL/zdxGstX5TJo0Ebc3KpePxYePoaUlLNMnjKGU7FxvP3GYxgMhlJhpIAAX1JT42wqm5Z2icrerphMIhwFQRAEQSh/3FWFs3DxOpb+lsyAAXNuEo55eHhU58EH5/O/H2L5bXnp2dUoom0jYmJ+x2zOKbLsyZMriWhbX7yOgiAIgiCIeLwTcs1m3ps8n65dJ2Nn53TLcvb2rnTp8h7vTZ5XatZMrFO7Oi3C3di37/tCy2VmXiE6ehJjxzwgPUsQBEEQBBGPd8LZswlcTHDFz69hkWWrVQvnxMkcEi5dKRVGMhjgoynjObD/TY4cWVJgmbS0RBYsGMK4sa1p0qSO9CxBEARBEMoldy3m8cyZeCpVqoPBULReNRrtcHauTkLCZfx8PUuFoQKr+7Fi2SSGjXyNPXu+pXHjR/DxaUBGxhViYlbzx9FZvDixD0/94wEM0q+EIrBYLBw/dYqcnBybP+Pr7Y1P5cpiPEEQBKFiiEcPD1fS0i7ZXD4zMwUXZ6dSZazgYH+i1n7MylU7+OXX/3H23CWcnBzo1+c+Bn/7Pv7+PtKjBJvIys6m58MPc5+LCyYbEsPOpKTQ/5FHeGXCBDGeIAiCUDHEY+3aAeTmniQz8wqOjpUKLZuaehFnp4v4B5Q+MWY0Grm/e0vu795Seo9wR1RycGD+gw/iZFf01/C7vXs5LSYTBEEQSoMWulsXcnVxYkD/cLZt+7TIslu2TGXIQxE4OthLCwmCIAiCIFRE8Qjw8ovDSLg4m0OHFtyihIV9+2aTnvYTzz4zWFpHEARBEAShlHFXFwn3ruzBkoVvMHTYG5w4sZzw8DF4e4cAFhISjrB9+6d4uB9g0Y9v4OHhKq0jCIIgCIJQkcUjQHAtf9atmsrceSuZv+ApNmy4AhYLdep48/yzXRj84EgcZLpaEARBEARBxGMeTk4OjBrRm1EjeksL/MVkZ+dwMuY8cfFJVPf3pUaNqtjZmcQwgiAIgiCUHfEo/PVkZmbx0acL+ebbVWRm+uPi4ktaWjxurhf5x5hePDaqj3h4BUEQBEEQ8SjAlSvXGDD4VVKuhtO79zrc3f31EQvJyaf44qvXWL7iNebO/jeurs5iMEEQBEEQbMYoJihfWCww5qkPwPAAvXpNtxKOAAY8PWvRv/+3JF5uy8SXPrfpnOnpmXwx42e69XyW4JDh1Kk/nL4PvMic/60kJydXjC4IgiAIIh6FssrmLQfYtj2LiIiJGG6xc4nBYKRTpzf4eekpDh+JLfR8R47G0qb9OL6amUrtOjMY+shBHhpykCpVJ/Pu5BN0uf+fxMUnieEFQRAEoYIg09bljEVL1tOkyWiMxsKb1s7OidDQYSz5OYoG9WsUWObixcv0G/g6zZt/TN26vW46Vr16G6pXb83u3TPpN+BfrFkxBVdXJ2mAYpCSmcmTS5diMhb9DHcsMZFudeuK0QRBEAQRj0LJcuhwLLWCW9hU1t+/OXv3Lb3l8dfe/IZatcb9STjewECzZqNYfvEAH3+6gJdeeFQawEbs7eyY9dlnZGVl2fyZmoGBYjhBEARBxKNQshgAi8VsU1mLxXLLqe2Uq2n8uuwgjw77vsjztGnzLHP+14Xnn3tYlgGyEaPRSJvwcDFEKWdtdDQZGRk2lw+qXp3QkBAxnCAIIh6FskPD0GCOHNuGn1/DIsueP7+dLp0Kngo9fCQWV7dGODi4FXkeD48ArqV6EX/xMgH+PtIIQrnAbLEw8pln6FGtGvamoh+KjiclERwRwbT//EeMJwiCiEeh7DBwQEcGD/2aJk2GYTTeeh3H7Ow0Dh6cxSdTXijweOq1dFxcKtt4VQNGowtZWdnSAEK5wmQ08l63bng4OhZZdt6BA0SJyQRBqABItnU5o2WLBnRs78H69W/dcvrabM5l9eqXGDSwHiF1C46jCwz0IzHxqE3XzM3NIjf3Ip6V3KQBBEEQBEHEo1DW+PSjCbi6rOLnn0eTnBxjdcRCYuJRFi16mKDA/fznrSdueY6aNariWSmRS5eOFHm9M2c206C+G15e7mJ8QRAEQSjnyLR1OcTd3YUlC97my69+YcbXvUi67I6zsy/p6Rep4pfBcxP68OjQf2Bvf+vmt7e347l/DuDtSf/H4MGLMJkcCiyXlXWNNWsm8u1XQ8XwgiAIgiDiUSirODjYM27sAJ58oh8XLiQSF5+IfzUfqlb1xmSyzeE89OFuRG8+wOLFj9Kz56e4uvrddDw5OZZffhnNE6Mb07lTMzG6IAiCIIh4FMp8A9uZCAz0IzDQr9ifNRmNfDJ1Ap9+tpAPprTGr0pXqlULx2LJ5ezZzVy7Gs1r/x7CsEfuv+WSP4IgCIIgiHgUKpj4nDB+MCOH92b12p2cOLEXDBB6X2M6dxyOs7OjGEkQBEEQRDwKws1UquTKgP4dxBCCIAiCUMGRbGtBEARBEARBxKMgCIIgCIJQ8si0tSAIQgEYAIvBwIMLFmAyFv2cHX/tGu169BDDCYIg4lEQBKFCikeDgXULF5Kbm2vzZ9zdZJclQRBEPAqCIFRYggICxAiCIAj5kJhHQRAEQRAEQcSjIAiCIAiCIOJREARBEARBEPEoCIIgCIIglAWKTJj5fu4Ktu88IpYSSj1paRlYLBZWR0XxyLhxYhChTHA+Lk6MIAhCmcJQyLGewDIxkSAIwl/ObiBMzCAIQlng/wENfpNj8E7ZjAAAAABJRU5ErkJggg=="}}},{"cell_type":"code","source":"test_df.head()","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:35:12.434880Z","iopub.execute_input":"2022-01-06T17:35:12.435280Z","iopub.status.idle":"2022-01-06T17:35:12.447459Z","shell.execute_reply.started":"2022-01-06T17:35:12.435243Z","shell.execute_reply":"2022-01-06T17:35:12.446652Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Để đánh giá được 1 câu có chân thành hay không, ta cần tìm ra được đặc trưng của nó cách đơn giản nhất đó là chúng ta cùng đi thống kê xem từ nào hay xuất hiện trong các câu chân thành và không chân thành nhất\n","metadata":{}},{"cell_type":"code","source":"class cout_Vocab(object):\n    def __init__(self):\n        self.vocab = {}\n        self.STOPWORDS = set()\n        self.STOPWORDS = set(stopwords.words('english'))\n    def build_vocab(self, lines):\n        for line in lines: \n            for word in line.split(' '):\n                word = word.lower()\n                if(word in self.STOPWORDS):\n                    continue\n                if(word not in self.vocab):\n                    self.vocab[word] = 0\n                self.vocab[word] += 1\n                ","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:35:12.448628Z","iopub.execute_input":"2022-01-06T17:35:12.449233Z","iopub.status.idle":"2022-01-06T17:35:12.456143Z","shell.execute_reply.started":"2022-01-06T17:35:12.449193Z","shell.execute_reply":"2022-01-06T17:35:12.455365Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Thống kê số lượng từ hay xuất hiện trong câu chân thành\nsincere_vocab = cout_Vocab()\nsincere_vocab.build_vocab(train_df[train_df['target']== 0]['question_text'])\nsincere_vocabulary = sorted(sincere_vocab.vocab.items(), reverse=True, key=lambda kv: kv[1])\n# Dùng sns vẽ biểu đồ 10 từ có tần suất xuất hiện cao nhất\nfor word, count in sincere_vocabulary[:10]:\n    print(word, count)\ndf_sincere_vocab = pd.DataFrame(sincere_vocabulary,columns = ['word_sincere', 'frequency'])\nsns.barplot(x = 'word_sincere', y = 'frequency', data = df_sincere_vocab[:10])","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:35:12.457348Z","iopub.execute_input":"2022-01-06T17:35:12.457594Z","iopub.status.idle":"2022-01-06T17:35:19.982391Z","shell.execute_reply.started":"2022-01-06T17:35:12.457561Z","shell.execute_reply":"2022-01-06T17:35:19.981705Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# thống kê số lượng từ hay xuất hiện trong câu không chân thành\ninsincere_vocab = cout_Vocab()\ninsincere_vocab.build_vocab(train_df[train_df['target']== 1]['question_text'])\ninsincere_vocabulary = sorted(insincere_vocab.vocab.items(), reverse=True, key=lambda kv: kv[1])\nfor word, count in insincere_vocabulary[:10]:\n    print(word, count)\ndf_insincere_vocab = pd.DataFrame(insincere_vocabulary,columns = ['word_insincere', 'frequency'])\nsns.barplot(y= 'word_insincere', x = 'frequency', data = df_insincere_vocab[5:25])","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:35:19.983585Z","iopub.execute_input":"2022-01-06T17:35:19.983838Z","iopub.status.idle":"2022-01-06T17:35:20.980914Z","shell.execute_reply.started":"2022-01-06T17:35:19.983800Z","shell.execute_reply":"2022-01-06T17:35:20.980103Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"**Nhận xét**: ở các câu không chân thành có sự xuất hiện đáng kể của các từ như white-black, sex-girl, americians-india-chinese, trump qua đó cũng có thể thấy phần lớp sẽ tập trung vào các loại như từ ngữ phản cảm, nội dung khiêu dâm, phân biệt chủng tộc, nhân vật chính trị","metadata":{}},{"cell_type":"code","source":"from wordcloud import WordCloud, STOPWORDS\nprint('Ảnh word cloud được tạo từ những câu hỏi chân thành:')\nsincere_wordcloud = WordCloud(width=600, height=400, background_color ='black', min_font_size = 10).generate(str(train_df[train_df['target'] == 0][\"question_text\"]))\n#Positive Word cloud\nplt.figure(figsize=(15,6), facecolor=None)\nplt.imshow(sincere_wordcloud)\nplt.axis(\"off\")\nplt.tight_layout(pad=0)\nplt.show();","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:35:20.982255Z","iopub.execute_input":"2022-01-06T17:35:20.982571Z","iopub.status.idle":"2022-01-06T17:35:21.660251Z","shell.execute_reply.started":"2022-01-06T17:35:20.982533Z","shell.execute_reply":"2022-01-06T17:35:21.658735Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print('Ảnh word cloud được tạo từ những câu hỏi thiếu chân thành:')\ninsincere_wordcloud = WordCloud(width=600, height=400, background_color ='white', min_font_size = 10).generate(str(train_df[train_df['target'] == 1][\"question_text\"]))\n#Positive Word cloud\nplt.figure(figsize=(15,6), facecolor=None)\nplt.imshow(insincere_wordcloud)\nplt.axis(\"off\")\nplt.tight_layout(pad=0)\nplt.show();","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:35:21.661277Z","iopub.execute_input":"2022-01-06T17:35:21.661524Z","iopub.status.idle":"2022-01-06T17:35:22.173954Z","shell.execute_reply.started":"2022-01-06T17:35:21.661496Z","shell.execute_reply":"2022-01-06T17:35:22.173296Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Xử lí dữ liệu","metadata":{}},{"cell_type":"markdown","source":"Qua bước đánh giá, ta thấy dữ liệu vẫn còn khá nhiều phức tạp cần xử lí loại bỏ bớt thông tin dư thừa như công thức toán học, đường dẫn, sửa lỗi chính tả, chuẩn hóa các từ đưa hết về chữ thường, thay đổi hết các từ viết tắt, xóa bỏ các kí tự đặc biệt. Ở đây em có thử xử lí loại bỏ stopword nhưng điều này vô tình làm thay đổi nghĩa của câu-> giảm độ chính xác của mô hình nên phần này em để nguyên stopword","metadata":{}},{"cell_type":"markdown","source":"# Loại bỏ các kí tự đặc biệt\nCác ký tự đặc biệt thường không mang nhiều về ý nghĩa giữa câu chân thành và không chân thành. Điều cần làm là loại bỏ chúng đi","metadata":{}},{"cell_type":"code","source":"punctuation_list =[',', '.', '\"', ':', ')', '(', '-', '!', '?', '|', ';', \"'\", '$', '&', '/', '[', ']', '>', '%', '=', '#', '*', '+', '\\\\', \n        '•', '~', '@', '£', '·', '_', '{', '}', '©', '^', '®', '`', '<', '→', '°', '€', '™', '›', '♥', '←', '×', '§', '″', '′', \n        '█', '…', '“', '★', '”', '–', '●', '►', '−', '¢', '¬', '░', '¡', '¶', '↑', '±', '¿', '▾', '═', '¦', '║', '―', '¥', '▓', \n        '—', '‹', '─', '▒', '：', '⊕', '▼', '▪', '†', '■', '’', '▀', '¨', '▄', '♫', '☆', '¯', '♦', '¤', '▲', '¸', '⋅', '‘', '∞', \n        '∙', '）', '↓', '、', '│', '（', '»', '，', '♪', '╩', '╚', '・', '╦', '╣', '╔', '╗', '▬', '❤', '≤', '‡', '√', '◄', '━', \n        '⇒', '▶', '≥', '╝', '♡', '◊', '。', '✈', '≡', '☺', '✔', '↵', '≈', '✓', '♣', '☎', '℃', '◦', '└', '‟', '～', '！', '○', \n        '◆', '№', '♠', '▌', '✿', '▸', '⁄', '□', '❖', '✦', '．', '÷', '｜', '┃', '／', '￥', '╠', '↩', '✭', '▐', '☼', '☻', '┐', \n        '├', '«', '∼', '┌', '℉', '☮', '฿', '≦', '♬', '✧', '〉', '－', '⌂', '✖', '･', '◕', '※', '‖', '◀', '‰', '\\x97', '↺', \n        '∆', '┘', '┬', '╬', '،', '⌘', '⊂', '＞', '〈', '⎙', '？', '☠', '⇐', '▫', '∗', '∈', '≠', '♀', '♔', '˚', '℗', '┗', '＊', \n        '┼', '❀', '＆', '∩', '♂', '‿', '∑', '‣', '➜', '┛', '⇓', '☯', '⊖', '☀', '┳', '；', '∇', '⇑', '✰', '◇', '♯', '☞', '´', \n        '↔', '┏', '｡', '◘', '∂', '✌', '♭', '┣', '┴', '┓', '✨', '\\xa0', '˜', '❥', '┫', '℠', '✒', '［', '∫', '\\x93', '≧', '］', \n        '\\x94', '∀', '♛', '\\x96', '∨', '◎', '↻', '⇩', '＜', '≫', '✩', '✪', '♕', '؟', '₤', '☛', '╮', '␊', '＋', '┈', '％', \n        '╋', '▽', '⇨', '┻', '⊗', '￡', '।', '▂', '✯', '▇', '＿', '➤', '✞', '＝', '▷', '△', '◙', '▅', '✝', '∧', '␉', '☭', \n        '┊', '╯', '☾', '➔', '∴', '\\x92', '▃', '↳', '＾', '׳', '➢', '╭', '➡', '＠', '⊙', '☢', '˝', '∏', '„', '∥', '❝', '☐', \n        '▆', '╱', '⋙', '๏', '☁', '⇔', '▔', '\\x91', '➚', '◡', '╰', '\\x85', '♢', '˙', '۞', '✘', '✮', '☑', '⋆', 'ⓘ', '❒', \n        '☣', '✉', '⌊', '➠', '∣', '❑', '◢', 'ⓒ', '\\x80', '〒', '∕', '▮', '⦿', '✫', '✚', '⋯', '♩', '☂', '❞', '‗', '܂', '☜', \n        '‾', '✜', '╲', '∘', '⟩', '＼', '⟨', '·', '✗', '♚', '∅', 'ⓔ', '◣', '͡', '‛', '❦', '◠', '✄', '❄', '∃', '␣', '≪', '｢', \n        '≅', '◯', '☽', '∎', '｣', '❧', '̅', 'ⓐ', '↘', '⚓', '▣', '˘', '∪', '⇢', '✍', '⊥', '＃', '⎯', '↠', '۩', '☰', '◥', \n        '⊆', '✽', '⚡', '↪', '❁', '☹', '◼', '☃', '◤', '❏', 'ⓢ', '⊱', '➝', '̣', '✡', '∠', '｀', '▴', '┤', '∝', '♏', 'ⓐ', \n        '✎', ';', '␤', '＇', '❣', '✂', '✤', 'ⓞ', '☪', '✴', '⌒', '˛', '♒', '＄', '✶', '▻', 'ⓔ', '◌', '◈', '❚', '❂', '￦', \n        '◉', '╜', '̃', '✱', '╖', '❉', 'ⓡ', '↗', 'ⓣ', '♻', '➽', '׀', '✲', '✬', '☉', '▉', '≒', '☥', '⌐', '♨', '✕', 'ⓝ', \n        '⊰', '❘', '＂', '⇧', '̵', '➪', '▁', '▏', '⊃', 'ⓛ', '‚', '♰', '́', '✏', '⏑', '̶', 'ⓢ', '⩾', '￠', '❍', '≃', '⋰', '♋', \n        '､', '̂', '❋', '✳', 'ⓤ', '╤', '▕', '⌣', '✸', '℮', '⁺', '▨', '╨', 'ⓥ', '♈', '❃', '☝', '✻', '⊇', '≻', '♘', '♞', \n        '◂', '✟', '⌠', '✠', '☚', '✥', '❊', 'ⓒ', '⌈', '❅', 'ⓡ', '♧', 'ⓞ', '▭', '❱', 'ⓣ', '∟', '☕', '♺', '∵', '⍝', 'ⓑ', \n        '✵', '✣', '٭', '♆', 'ⓘ', '∶', '⚜', '◞', '்', '✹', '➥', '↕', '̳', '∷', '✋', '➧', '∋', '̿', 'ͧ', '┅', '⥤', '⬆', '⋱', \n        '☄', '↖', '⋮', '۔', '♌', 'ⓛ', '╕', '♓', '❯', '♍', '▋', '✺', '⭐', '✾', '♊', '➣', '▿', 'ⓑ', '♉', '⏠', '◾', '▹', \n        '⩽', '↦', '╥', '⍵', '⌋', '։', '➨', '∮', '⇥', 'ⓗ', 'ⓓ', '⁻', '⎝', '⌥', '⌉', '◔', '◑', '✼', '♎', '♐', '╪', '⊚', \n        '☒', '⇤', 'ⓜ', '⎠', '◐', '⚠', '╞', '◗', '⎕', 'ⓨ', '☟', 'ⓟ', '♟', '❈', '↬', 'ⓓ', '◻', '♮', '❙', '♤', '∉', '؛', \n        '⁂', 'ⓝ', '־', '♑', '╫', '╓', '╳', '⬅', '☔', '☸', '┄', '╧', '׃', '⎢', '❆', '⋄', '⚫', '̏', '☏', '➞', '͂', '␙', \n        'ⓤ', '◟', '̊', '⚐', '✙', '↙', '̾', '℘', '✷', '⍺', '❌', '⊢', '▵', '✅', 'ⓖ', '☨', '▰', '╡', 'ⓜ', '☤', '∽', '╘', \n        '˹', '↨', '♙', '⬇', '♱', '⌡', '⠀', '╛', '❕', '┉', 'ⓟ', '̀', '♖', 'ⓚ', '┆', '⎜', '◜', '⚾', '⤴', '✇', '╟', '⎛', \n        '☩', '➲', '➟', 'ⓥ', 'ⓗ', '⏝', '◃', '╢', '↯', '✆', '˃', '⍴', '❇', '⚽', '╒', '̸', '♜', '☓', '➳', '⇄', '☬', '⚑', \n        '✐', '⌃', '◅', '▢', '❐', '∊', '☈', '॥', '⎮', '▩', 'ு', '⊹', '‵', '␔', '☊', '➸', '̌', '☿', '⇉', '⊳', '╙', 'ⓦ', \n        '⇣', '｛', '̄', '↝', '⎟', '▍', '❗', '״', '΄', '▞', '◁', '⛄', '⇝', '⎪', '♁', '⇠', '☇', '✊', 'ி', '｝', '⭕', '➘', \n        '⁀', '☙', '❛', '❓', '⟲', '⇀', '≲', 'ⓕ', '⎥', '\\u06dd', 'ͤ', '₋', '̱', '̎', '♝', '≳', '▙', '➭', '܀', 'ⓖ', '⇛', '▊', \n        '⇗', '̷', '⇱', '℅', 'ⓧ', '⚛', '̐', '̕', '⇌', '␀', '≌', 'ⓦ', '⊤', '̓', '☦', 'ⓕ', '▜', '➙', 'ⓨ', '⌨', '◮', '☷', \n        '◍', 'ⓚ', '≔', '⏩', '⍳', '℞', '┋', '˻', '▚', '≺', 'ْ', '▟', '➻', '̪', '⏪', '̉', '⎞', '┇', '⍟', '⇪', '▎', '⇦', '␝', \n        '⤷', '≖', '⟶', '♗', '̴', '♄', 'ͨ', '̈', '❜', '̡', '▛', '✁', '➩', 'ா', '˂', '↥', '⏎', '⎷', '̲', '➖', '↲', '⩵', '̗', '❢', \n        '≎', '⚔', '⇇', '̑', '⊿', '̖', '☍', '➹', '⥊', '⁁', '✢']","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:35:22.177070Z","iopub.execute_input":"2022-01-06T17:35:22.177445Z","iopub.status.idle":"2022-01-06T17:35:22.212443Z","shell.execute_reply.started":"2022-01-06T17:35:22.177408Z","shell.execute_reply":"2022-01-06T17:35:22.211807Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def remove_punctuation(text):\n    for punctuation in punctuation_list:\n        if punctuation in text:\n            text = text.replace(punctuation, '{}' .format(punctuation))\n    return text","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:35:22.213723Z","iopub.execute_input":"2022-01-06T17:35:22.214047Z","iopub.status.idle":"2022-01-06T17:35:22.226312Z","shell.execute_reply.started":"2022-01-06T17:35:22.214012Z","shell.execute_reply":"2022-01-06T17:35:22.225526Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Loại bỏ số\nTa có thể thấy các con số không mang ý nghĩa về câu chân thành hay không chân thành. Vì vậy việc đơn giản ta cần làm là loại bỏ những thứ vô nghĩa để mô hình chính xác hơn","metadata":{}},{"cell_type":"code","source":"def clean_numbers(text):\n    if bool(re.search(r'\\d', text)):\n        text = re.sub('[0-9]{5,}', '#####', text)\n        text = re.sub('[0-9]{4}', '####', text)\n        text = re.sub('[0-9]{3}', '###', text)\n        text = re.sub('[0-9]{2}', '##', text)\n    return text","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:35:22.227740Z","iopub.execute_input":"2022-01-06T17:35:22.228253Z","iopub.status.idle":"2022-01-06T17:35:22.235217Z","shell.execute_reply.started":"2022-01-06T17:35:22.228217Z","shell.execute_reply":"2022-01-06T17:35:22.234415Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Sửa lỗi chính tả\nLỗi này xuất hiện khi người dùng sử dụng lẫn lộn giữa Anh-Anh và Anh-Mỹ hay gõ phím sai. \nCách xử lí: em sử dụng 1 mảng sẵn mispell_dict convert Anh-Anh sang Anh-Mỹ.Mục đích của việc convert này nhằm đạt được độ phủ cao với vocab được build từ các tệp trong embedding.zip  ","metadata":{}},{"cell_type":"code","source":"mispell_dict = {'colour': 'color', 'centre': 'center', 'favourite': 'favorite', 'travelling': 'traveling', 'counselling': 'counseling', 'theatre': 'theater', 'cancelled': 'canceled', 'labour': 'labor', 'organisation': 'organization', 'wwii': 'world war 2', 'citicise': 'criticize', 'youtu ': 'youtube ', 'Qoura': 'Quora', 'sallary': 'salary', 'Whta': 'What', 'narcisist': 'narcissist', 'howdo': 'how do', 'whatare': 'what are', 'howcan': 'how can', 'howmuch': 'how much', 'howmany': 'how many', 'whydo': 'why do', 'doI': 'do I', 'theBest': 'the best', 'howdoes': 'how does', 'mastrubation': 'masturbation', 'mastrubate': 'masturbate', \"mastrubating\": 'masturbating', 'pennis': 'penis', 'Etherium': 'bitcoin', 'narcissit': 'narcissist', 'bigdata': 'big data', '2k17': '2017', '2k18': '2018', 'qouta': 'quota', 'exboyfriend': 'ex boyfriend', 'airhostess': 'air hostess', \"whst\": 'what', 'watsapp': 'whatsapp', 'demonitisation': 'demonetization', 'demonitization': 'demonetization', 'demonetisation': 'demonetization', \n                'electroneum':'bitcoin','nanodegree':'degree','hotstar':'star','dream11':'dream','ftre':'fire','tensorflow':'framework','unocoin':'bitcoin',\n                'lnmiit':'limit','unacademy':'academy','altcoin':'bitcoin','altcoins':'bitcoin','litecoin':'bitcoin','coinbase':'bitcoin','cryptocurency':'cryptocurrency',\n                'simpliv':'simple','quoras':'quora','schizoids':'psychopath','remainers':'remainder','twinflame':'soulmate','quorans':'quora','brexit':'demonetized',\n                'iiest':'institute','dceu':'comics','pessat':'exam','uceed':'college','bhakts':'devotee','boruto':'anime',\n                'cryptocoin':'bitcoin','blockchains':'blockchain','fiancee':'fiance','redmi':'smartphone','oneplus':'smartphone','qoura':'quora','deepmind':'framework','ryzen':'cpu','whattsapp':'whatsapp',\n                'undertale':'adventure','zenfone':'smartphone','cryptocurencies':'cryptocurrencies','koinex':'bitcoin','zebpay':'bitcoin','binance':'bitcoin','whtsapp':'whatsapp',\n                'reactjs':'framework','bittrex':'bitcoin','bitconnect':'bitcoin','bitfinex':'bitcoin','yourquote':'your quote','whyis':'why is','jiophone':'smartphone',\n                'dogecoin':'bitcoin','onecoin':'bitcoin','poloniex':'bitcoin','7700k':'cpu','angular2':'framework','segwit2x':'bitcoin','hashflare':'bitcoin','940mx':'gpu',\n                'openai':'framework','hashflare':'bitcoin','1050ti':'gpu','nearbuy':'near buy','freebitco':'bitcoin','antminer':'bitcoin','filecoin':'bitcoin','whatapp':'whatsapp',\n                'empowr':'empower','1080ti':'gpu','crytocurrency':'cryptocurrency','8700k':'cpu','whatsaap':'whatsapp','g4560':'cpu','payymoney':'pay money',\n                'fuckboys':'fuck boys','intenship':'internship','zcash':'bitcoin','demonatisation':'demonetization','narcicist':'narcissist','mastuburation':'masturbation',\n                'trignometric':'trigonometric','cryptocurreny':'cryptocurrency','howdid':'how did','crytocurrencies':'cryptocurrencies','phycopath':'psychopath',\n                'bytecoin':'bitcoin','possesiveness':'possessiveness','scollege':'college','humanties':'humanities','altacoin':'bitcoin','demonitised':'demonetized',\n                'brasília':'brazilia','accolite':'accolyte','econimics':'economics','varrier':'warrier','quroa':'quora','statergy':'strategy','langague':'language',\n                'splatoon':'game','7600k':'cpu','gate2018':'gate 2018','in2018':'in 2018','narcassist':'narcissist','jiocoin':'bitcoin','hnlu':'hulu','7300hq':'cpu',\n                'weatern':'western','interledger':'blockchain','deplation':'deflation', 'cryptocurrencies':'cryptocurrency', 'bitcoin':'blockchain cryptocurrency',}","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:35:22.236652Z","iopub.execute_input":"2022-01-06T17:35:22.237171Z","iopub.status.idle":"2022-01-06T17:35:22.253580Z","shell.execute_reply.started":"2022-01-06T17:35:22.237134Z","shell.execute_reply":"2022-01-06T17:35:22.252788Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import re\ndef get_misspelled_dict_and_regex(mispell_dict):\n    mispell_re = re.compile('(%s)' % '|'.join(mispell_dict.keys()))\n    return mispell_dict, mispell_re\n\nmispellings, mispellings_re = get_misspelled_dict_and_regex(mispell_dict)\ndef replace_typical_misspell(text):\n    def replace(match):\n        return mispellings[match.group(0)]\n    return mispellings_re.sub(replace, text)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:35:22.256365Z","iopub.execute_input":"2022-01-06T17:35:22.256555Z","iopub.status.idle":"2022-01-06T17:35:22.268491Z","shell.execute_reply.started":"2022-01-06T17:35:22.256533Z","shell.execute_reply":"2022-01-06T17:35:22.267769Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Đổi tất cả các chữ viết tắt về viết bình thường\nViệc này cũng giúp cho đạt được độ phủ cao với vocab được build từ tệp trong embedding.zip","metadata":{}},{"cell_type":"code","source":"contraction_dict = {\n    \"ain't\": \"is not\", \n    \"aren't\": \"are not\",\n    \"can't\": \"cannot\", \n    \"'cause\": \"because\", \n    \"could've\": \"could have\", \n    \"couldn't\": \"could not\", \n    \"didn't\": \"did not\",  \n    \"doesn't\": \"does not\", \n    \"don't\": \"do not\", \n    \"hadn't\": \"had not\", \n    \"hasn't\": \"has not\", \n    \"haven't\": \"have not\", \n    \"he'd\": \"he would\",\n    \"he'll\": \"he will\", \n    \"he's\": \"he is\", \n    \"how'd\": \"how did\", \n    \"how'd'y\": \"how do you\", \n    \"how'll\": \"how will\", \n    \"how's\": \"how is\",  \n    \"I'd\": \"I would\", \n    \"I'd've\": \"I would have\",\n    \"I'll\": \"I will\", \n    \"I'll've\": \"I will have\",\n    \"I'm\": \"I am\", \n    \"I've\": \"I have\", \n    \"i'd\": \"i would\", \n    \"i'd've\": \"i would have\", \n    \"i'll\": \"i will\",  \n    \"i'll've\": \"i will have\",\n    \"i'm\": \"i am\", \n    \"i've\": \"i have\", \n    \"isn't\": \"is not\", \n    \"it'd\": \"it would\", \n    \"it'd've\": \"it would have\", \n    \"it'll\": \"it will\", \n    \"it'll've\": \"it will have\",\n    \"it's\": \"it is\", \n    \"let's\": \"let us\", \n    \"ma'am\": \"madam\", \n    \"mayn't\": \"may not\", \n    \"might've\": \"might have\",\n    \"mightn't\": \"might not\",\n    \"mightn't've\": \"might not have\", \n    \"must've\": \"must have\", \n    \"mustn't\": \"must not\", \n    \"mustn't've\": \"must not have\", \n    \"needn't\": \"need not\", \n    \"needn't've\": \"need not have\",\n    \"o'clock\": \"of the clock\", \n    \"oughtn't\": \"ought not\", \n    \"oughtn't've\": \"ought not have\", \n    \"shan't\": \"shall not\", \n    \"sha'n't\": \"shall not\", \"shan't've\": \"shall not have\", \n    \"she'd\": \"she would\", \"she'd've\": \"she would have\", \n    \"she'll\": \"she will\", \"she'll've\": \"she will have\", \n    \"she's\": \"she is\", \"should've\": \"should have\", \n    \"shouldn't\": \"should not\", \"shouldn't've\": \"should not have\", \n    \"so've\": \"so have\",\"so's\": \"so as\", \"this's\": \"this is\",\n    \"that'd\": \"that would\", \"that'd've\": \"that would have\", \n    \"that's\": \"that is\", \"there'd\": \"there would\", \n    \"there'd've\": \"there would have\", \"there's\": \"there is\", \n    \"here's\": \"here is\",\"they'd\": \"they would\", \"they'd've\": \"they would have\", \n    \"they'll\": \"they will\", \"they'll've\": \"they will have\", \n    \"they're\": \"they are\", \"they've\": \"they have\", \n    \"to've\": \"to have\", \"wasn't\": \"was not\", \n    \"we'd\": \"we would\", \"we'd've\": \"we would have\", \n    \"we'll\": \"we will\", \"we'll've\": \"we will have\", \n    \"we're\": \"we are\", \"we've\": \"we have\", \n    \"weren't\": \"were not\", \"what'll\": \"what will\", \n    \"what'll've\": \"what will have\", \"what're\": \"what are\",  \n    \"what's\": \"what is\", \"what've\": \"what have\", \"when's\": \"when is\", \n    \"when've\": \"when have\", \"where'd\": \"where did\", \"where's\": \"where is\", \n    \"where've\": \"where have\", \"who'll\": \"who will\", \"who'll've\": \"who will have\", \n    \"who's\": \"who is\", \"who've\": \"who have\", \"why's\": \"why is\", \"why've\": \"why have\", \n    \"will've\": \"will have\", \"won't\": \"will not\", \"won't've\": \"will not have\", \n    \"would've\": \"would have\", \"wouldn't\": \"would not\", \"wouldn't've\": \"would not have\", \n    \"y'all\": \"you all\", \"y'all'd\": \"you all would\",\"y'all'd've\": \"you all would have\",\n    \"y'all're\": \"you all are\",\"y'all've\": \"you all have\",\"you'd\": \"you would\", \n    \"you'd've\": \"you would have\", \"you'll\": \"you will\", \"you'll've\": \"you will have\", \n    \"you're\": \"you are\", \"you've\": \"you have\"}","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:35:22.269777Z","iopub.execute_input":"2022-01-06T17:35:22.270376Z","iopub.status.idle":"2022-01-06T17:35:22.285554Z","shell.execute_reply.started":"2022-01-06T17:35:22.270249Z","shell.execute_reply":"2022-01-06T17:35:22.284760Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def get_contractions_dict_and_regex(contraction_dict):\n    contraction_re = re.compile('(%s)' % '|'.join(contraction_dict.keys()))\n    return contraction_dict, contraction_re\n\ncontractions, contractions_re = get_contractions_dict_and_regex(contraction_dict)\n\ndef replace_contractions(text):\n    def replace(match):\n        return contractions[match.group(0)]\n    return contractions_re.sub(replace, text)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:35:22.286540Z","iopub.execute_input":"2022-01-06T17:35:22.287030Z","iopub.status.idle":"2022-01-06T17:35:22.299105Z","shell.execute_reply.started":"2022-01-06T17:35:22.286994Z","shell.execute_reply":"2022-01-06T17:35:22.298358Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Loại bỏ stopword: những từ thường xuất hiện(không dùng)","metadata":{}},{"cell_type":"code","source":"# import nltk\n# from nltk.tokenize.toktok import ToktokTokenizer\n\n# stopword_list = nltk.corpus.stopwords.words('english')\n# def remove_stopwords(text, is_lower_case=True):\n#     tokenizer = ToktokTokenizer()\n#     tokens = tokenizer.tokenize(text)\n#     tokens = [token.strip() for token in tokens]\n#     if is_lower_case:\n#         filtered_tokens = [token for token in tokens if token not in stopword_list]\n#     else:\n#         filtered_tokens = [token for token in tokens if token.lower() not in stopword_list]\n#     filtered_text = ' '.join(filtered_tokens)\n#     return filtered_text","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:35:22.300408Z","iopub.execute_input":"2022-01-06T17:35:22.301090Z","iopub.status.idle":"2022-01-06T17:35:22.307375Z","shell.execute_reply.started":"2022-01-06T17:35:22.301054Z","shell.execute_reply":"2022-01-06T17:35:22.306704Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"**Xử lí chuẩn hóa chuỗi:**\n- đổi hết thành chữ thường\n- xóa hết kí tự đặc biệt\n- xóa hết số\n- sửa lỗi chính tả\n- sửa chữ viết tắt\n","metadata":{}},{"cell_type":"code","source":"def clean_questions(x):\n    x = x.lower()\n    x = remove_punctuation(x)\n    x = clean_numbers(x)\n    x = replace_typical_misspell(x)\n#     x = remove_stopwords(x)\n    x = replace_contractions(x)\n    x = x.replace(\"'\",\"\")\n    return x","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:35:22.308757Z","iopub.execute_input":"2022-01-06T17:35:22.309607Z","iopub.status.idle":"2022-01-06T17:35:22.317599Z","shell.execute_reply.started":"2022-01-06T17:35:22.309568Z","shell.execute_reply":"2022-01-06T17:35:22.316909Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"**Đưa nó vào xử lí ở cột question_text qua đó ta đã làm sach được phần nào của dữ liệu**","metadata":{}},{"cell_type":"code","source":"train_df['question_text'] = train_df['question_text'].apply(lambda x: clean_questions(x))\ntest_df['question_text'] = test_df['question_text'].apply(lambda x: clean_questions(x))","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:35:22.320496Z","iopub.execute_input":"2022-01-06T17:35:22.320703Z","iopub.status.idle":"2022-01-06T17:37:55.813778Z","shell.execute_reply.started":"2022-01-06T17:35:22.320681Z","shell.execute_reply":"2022-01-06T17:37:55.813039Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# lay ra 10% train de lam validate\nprint(train_df[[\"question_text\",\"target\"]]);\n\ntrain_df, val_df = train_test_split(train_df, test_size=0.1, random_state=2021)\n\n# some config values \nembed_size = 300 # Độ dài của mỗi vector từ\nmax_features = 50000 # Số lượng từ tối đa trong từ điển sẽ sử dụng\nmaxlen = 100 # Số lượng từ tối đa trong một câu\n\n# Điền \"na\" vào các dữ liệu còn trống \ntrain_X = train_df[\"question_text\"].fillna(\"_na_\").values\nval_X = val_df[\"question_text\"].fillna(\"_na_\").values\ntest_X = test_df[\"question_text\"].fillna(\"_na_\").values","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:37:55.815029Z","iopub.execute_input":"2022-01-06T17:37:55.815271Z","iopub.status.idle":"2022-01-06T17:37:56.678790Z","shell.execute_reply.started":"2022-01-06T17:37:55.815240Z","shell.execute_reply":"2022-01-06T17:37:56.678048Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"**Nhận xét:** Ta có question text là văn bản, nhưng máy tính không thể hiểu được văn bản để phân loại hay làm gì cả, để dễ huấn luyện mô hình ta cần chuyển nó sang các vector số nguyên. \n\nta sử dụng keras.Tokenizer để thực hiện tokenizer. tokenizer chia các văn bản thành từng chữ thống kê đánh số chúng. Từ word_index ánh xạ sang tập train và encode sang mã hóa onehot\n\none-hot vector: Đây là kỹ thuật biểu diễn từ bằng vector có số chiều bằng số từ vựng. Vector này có duy nhất một chiều có giá trị bằng 1 ứng với từ đang biểu diễn, các vị trí khác có giá trị 0. Ví dụ [1,0,0,0…0]. Biểu diễn này giải quyết được mẫu thuẫn tiềm năng của biểu diễn bằng số. Tuy nhiên, nhược điểm của phương pháp này là số chiều vector rất lớn, ảnh hưởng đến quá trình xử lý cũng như lưu trữ.","metadata":{}},{"cell_type":"code","source":"# mã hóa từ thành số, câu thành vector số \ntokenizer = Tokenizer(num_words=max_features)\ntokenizer.fit_on_texts(list(train_X))\nprint(\"Size of vocabulary: \", len(tokenizer.word_index))\nprint(\"20 từ đầu tiên trong từ điển:\")\nprint(list(tokenizer.word_index.items())[:20])\n\n","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:37:56.680698Z","iopub.execute_input":"2022-01-06T17:37:56.681127Z","iopub.status.idle":"2022-01-06T17:38:17.952212Z","shell.execute_reply.started":"2022-01-06T17:37:56.681090Z","shell.execute_reply":"2022-01-06T17:38:17.950709Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Sau khi có được vector one hot của các từ trong tập train nhiệm vụ của chúng ta là đưa các câu thành các vector thông qua texts_to_sequences. Chắc chắn không thể tránh khỏi việc các chuỗi có độ dài không bằng nhau nên việc cần phải làm là padding tất cả các chuỗi luôn có độ dài bằng 100 từ( nếu câu có ít hơn 100 từ thì điền 0 cho đủ, nếu nhiều hơn thì cắt bớt). khi đó ta sẽ được vector số 100x1","metadata":{}},{"cell_type":"code","source":"print(train_X[0])\ntrain_X = tokenizer.texts_to_sequences(train_X)\nval_X = tokenizer.texts_to_sequences(val_X)\ntest_X = tokenizer.texts_to_sequences(test_X)\n# Đảm bảo mỗi câu hỏi luôn dài 100 từ\ntrain_X = pad_sequences(train_X, maxlen=maxlen)\nval_X = pad_sequences(val_X, maxlen=maxlen)\ntest_X = pad_sequences(test_X, maxlen=maxlen)\nprint(train_X[0])","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:38:17.953624Z","iopub.execute_input":"2022-01-06T17:38:17.953886Z","iopub.status.idle":"2022-01-06T17:38:58.258917Z","shell.execute_reply.started":"2022-01-06T17:38:17.953837Z","shell.execute_reply":"2022-01-06T17:38:58.257252Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"## Get the target values\ntrain_y = train_df['target'].values\nval_y = val_df['target'].values","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:38:58.260335Z","iopub.execute_input":"2022-01-06T17:38:58.260589Z","iopub.status.idle":"2022-01-06T17:38:58.264800Z","shell.execute_reply.started":"2022-01-06T17:38:58.260556Z","shell.execute_reply":"2022-01-06T17:38:58.264148Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Mô hình\nBài toán này là 1 bài toán nhị phân. Để xử lí chúng ta có thể nghĩ đến một vài mô hình như Logistic Regression, SVM, LSTM,GRU. Ở đây em lựa chọn GRU \nKiến thức cần biết:\nMạng GRU là 1 phiên bản hoàn thiện của mạng RNN\nNói qua 1 chút về mạng RNN:\n![image.png](attachment:93bc2048-90dc-4b76-a209-2786c9c4edc4.png)\n\nMạng RNN là 1 mạng chuyên dùng để xử lí dữ liệu dạng chuỗi ví dụ như hoàn thành câu..... Tại sao lại như vậy? Nó sử dụng 1 bộ nhớ để lưu lại thông tin xử lí từ các bước trước để dựa vào đó đưa ra dự đoán chính xác cho bước hiện tại.Mạng đơn giản chỉ dùng 1 hàm tanh dựa vào input vào và history để tìm xác xuất thứ xuất hiện sau.\n\nThế nhưng cũng giống như các mạng neuron khác, nó sẽ gán các trong số đầu vào của nó và tạo ra đầu ra và 1 trong những vấn đề của mạng này đó chính là việc biến mất gradient\n\nĐiều này đã được mạng GRU cải thiện\nĐể giải quyết vấn đề biến mất gradient của mạng RNN, GRU đã được sử dụng thêm cổng cập nhật và công cài đặt lại(update gate và reset gate). Về cơ bản, đó chính là hai vector quyết định thông tin nào sẽ được truyền cho đầu ra. Điều đặc biệt là nó có thể được đào tạo để giữ thông tin từ lâu trước đó, không hề xóa thông tin không liên quan đến dự đoán đầu ra.\n\n![Image-from-38-The-GRU-model-combines-the-forget-gate-and-input-gate-into-a-single.png](attachment:7eaf2ee2-0dc4-4823-a781-56fc4ac98f9a.png)\n\nupdate gate: chịu trách nhiệm xác định lượng thông tin trước đó cần chuyển qua trạng thái tiếp theo. \n\ninput:  $x_t$\n\n$$z_t=\\sigma(W_z(x_t)+U_z(h_{t-1})$$ \nCả hai kết quả được cộng với nhau và qua hàm sigmoid để cho kết quả trong khoảng [0,1]. \n\nupdate gate giúp mô hình xác định được lượng thông tin trong quá khứ (thông tin ở bước t-1) cần chuyển đến tương lai (bước t). Điều này giúp mô hình có thể quyết định copy tất cả thông tin từ quá khứ và loại bỏ nguy cơ mất mát gradient. Bây giờ việc của bạn chỉ là nhớ history theo công thức $z_t$.\n\nreset gate: thiết lập lại quyết định lượng thông tin quá khứ cần thiết cần bỏ qua.\n\n$$r_t=\\sigma(W_r(x_t)+U_z(h_{t-1})$$ \n\n\nCurrent memory content (Nội dung nhớ hiện tại):xác định sự ảnh hưởng của các từ trên ảnh hưởng đến kết quả hay đầu ra hay không\n\nHãy xem chính xác các cổng nói trên có ảnh hưởng thế nào đến kết quả đầu ra nhé. Đầu tiên, tôi sẽ sử dụng cổng reset. \n\n\nCác bước thực hiện của công thức trên:\n\n- B1:$ x_t . W_r + h_{t-1} . U_r$\n- B2:$ r_t * Uh_{t-1}$  - Phép nhân chập sẽ xác định xem những gì cần phải xóa khỏi ở thời điểm trước đó.\n- Cộng kết quả ở bước 1 và 2\n- Sử dụng hàm tanh.=> output: $h'_t$\n\nFinal memory at current time step (bộ nhớ tại thời điểm hiện tại):output là lượng thông tin hữu ích đi ra từ h_t được đưa vào mạng để train tiếp\n\nỞ bước cuối cùng, đầu ra của mạng là cần tính thông tin hữu ích $h_t$ - vector chứa toàn bộ thông tin ở tại thời điểm t và truyền nó đi. Để thực hiện điều này, cần có cổng update. Nó xác định nội dung thu thập từ bộ nhớ hiện tại - $h'_t$ và những gì từ các bước trước đó $h_(t-1)$.\n\ninput: $h'_t, z_t$\n\nCác bước thực hiện như sau:\n\n- B1:$ z_t * h_{t-1}$\n- B2: $ (1-z_t) * h'_t$\n- Thực hiện phép cộng với kết quả ở bước 1 và 2.=>$h_t$-là thông tin hiện tại có ý nghĩa cần được lưu đưa vào mạng tiếp theo\nBằng cách này GRU có thể lưu trữ và lọc các thông tin thông qua 2 cổng qua đó giúp loại bỏ vấn đề gradient biến mất","metadata":{},"attachments":{"93bc2048-90dc-4b76-a209-2786c9c4edc4.png":{"image/png":"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"},"7eaf2ee2-0dc4-4823-a781-56fc4ac98f9a.png":{"image/png":"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"}}},{"cell_type":"code","source":"inp = Input(shape=(maxlen,))\nx = Embedding(max_features, embed_size)(inp)\nx = Bidirectional(GRU(64, return_sequences=True))(x)\nx = GlobalMaxPool1D()(x)\nx = Dense(16, activation=\"relu\")(x)\nx = Dropout(0.1)(x)\nx = Dense(1, activation=\"sigmoid\")(x)\nmodel = Model(inputs=inp, outputs=x)\nmodel.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])\n\nprint(model.summary())","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:38:58.266524Z","iopub.execute_input":"2022-01-06T17:38:58.267051Z","iopub.status.idle":"2022-01-06T17:39:01.358759Z","shell.execute_reply.started":"2022-01-06T17:38:58.267013Z","shell.execute_reply":"2022-01-06T17:39:01.357609Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Input: là các vector chuỗi tương ứng với câu hỏi, Một chuỗi có 100 từ tương đương vector 100 chiều\n\nembedding layer:  embedding sang một không gian mới có chiều nhỏ hơn, giảm chiều dữ liệu= max_feature*embed_size=5000*300\n\nTầng Bidirection sẽ giúp máy học được nghĩa của mỗi câu dựa trên thứ tự của các từ trên mạng noron.Các từ có nghĩa gần giống nhau thì sẽ có giá trị vector gần như nhau\n\nlớp dense : là lớp mạng neuron fully-connected -> cho qua hàm sigmoid -> dự đoán\n\nthuật toán tối ưu hóa: adam","metadata":{}},{"cell_type":"code","source":"model.fit(train_X, train_y, batch_size=512, epochs=2, validation_data=(val_X, val_y))","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:39:01.360621Z","iopub.execute_input":"2022-01-06T17:39:01.361020Z","iopub.status.idle":"2022-01-06T17:41:58.062279Z","shell.execute_reply.started":"2022-01-06T17:39:01.360981Z","shell.execute_reply":"2022-01-06T17:41:58.061595Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Threshold\nBình thường ở các mô hình phân lớp nhị phân kết quả dự đoán luôn thuộc (0,1) threshold=0.5.\n\nNếu threshold < 0.5 thì phân lớp là negative, ngược lại thì là positive\n\nNhưng trong một số trường hợp, giá trị 0.5 này có thể chưa phải là tốt nhất.\nTrong bài toán này, mục đích của chúng ta là giết nhầm còn hơn bỏ sót chính vì thế ta set threshold <0.5 (điều này đồng nghĩa với việc ta thà xác định nhầm 1 câu chân thành thành không chân thành còn hơn là bỏ lỡ chúng)\n","metadata":{}},{"cell_type":"code","source":"pred_noemb_val_y = model.predict([val_X], batch_size=1024, verbose=1)\nfor thresh in np.arange(0.1, 0.501, 0.01):\n    thresh = np.round(thresh, 2)\n    print(\"F1 score at threshold {0} is {1}\".format(thresh, metrics.f1_score(val_y, (pred_noemb_val_y>thresh).astype(int))))\n    ","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:41:58.063738Z","iopub.execute_input":"2022-01-06T17:41:58.064014Z","iopub.status.idle":"2022-01-06T17:42:02.931564Z","shell.execute_reply.started":"2022-01-06T17:41:58.063979Z","shell.execute_reply":"2022-01-06T17:42:02.930810Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"del model, inp, x\nimport gc; gc.collect()\ntime.sleep(10)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:42:02.933058Z","iopub.execute_input":"2022-01-06T17:42:02.933303Z","iopub.status.idle":"2022-01-06T17:42:13.174396Z","shell.execute_reply.started":"2022-01-06T17:42:02.933271Z","shell.execute_reply":"2022-01-06T17:42:13.173560Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Cải thiện hiệu suất mô hình với một số tool Embedding","metadata":{}},{"cell_type":"markdown","source":"**Xem xét các file nhúng được cung cấp. Chúng được để trong một file zip nên ta tiến hành unzip**\n\nVai trò của file embedding: mã hóa các word thành 1 vecto để model xử lí được. Thay vì ta thống kê chúng thì từ những file embedding này nó đã thực hiện sẵn và mã hóa các từ thành các vector số\nƯu điểm: do đã được train nên các vector này sẽ giúp những từ có gần nghĩa nhau thì sẽ đứng cạnh nhau\n\n![1_mWerYTuy9xH4SlRY9fFg1A.jpeg](attachment:bf4e9938-1011-442e-9907-f08cee04fb22.jpeg)","metadata":{},"attachments":{"bf4e9938-1011-442e-9907-f08cee04fb22.jpeg":{"image/jpeg":"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"}}},{"cell_type":"code","source":"from zipfile import ZipFile\nfile_name = \"../input/quora-insincere-questions-classification/embeddings.zip\"\nwith ZipFile(file_name, 'r') as zip:\n     # printing all the contents of the zip file\n    zip.printdir()\n  \n    # extracting all the files\n    print('Extracting all the files now...')\n    zip.extractall()\n    print('Done!')","metadata":{"execution":{"iopub.status.busy":"2022-01-07T00:25:33.086729Z","iopub.execute_input":"2022-01-07T00:25:33.086978Z","iopub.status.idle":"2022-01-07T00:29:06.269692Z","shell.execute_reply.started":"2022-01-07T00:25:33.086951Z","shell.execute_reply":"2022-01-07T00:29:06.268812Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Lưu lại đường dẫn file nhúng","metadata":{}},{"cell_type":"code","source":"glove = '../working/glove.840B.300d/glove.840B.300d.txt'\nparagram =  '../working/paragram_300_sl999/paragram_300_sl999.txt'\nwiki_news = '../working/wiki-news-300d-1M/wiki-news-300d-1M.vec'","metadata":{"execution":{"iopub.status.busy":"2022-01-07T00:29:06.278456Z","iopub.execute_input":"2022-01-07T00:29:06.278760Z","iopub.status.idle":"2022-01-07T00:29:09.462234Z","shell.execute_reply.started":"2022-01-07T00:29:06.278724Z","shell.execute_reply":"2022-01-07T00:29:09.437619Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Cài đặt load load_embed dùng để load 3 loại nhúng","metadata":{}},{"cell_type":"code","source":"def load_embed(file):\n    def get_coefs(word,*arr): \n        return word, np.asarray(arr, dtype='float32')\n    \n    if file == wiki_news:\n        embeddings_index = dict(get_coefs(*o.split(\" \")) for o in open(file) if len(o)>100)\n    else:\n        embeddings_index = dict(get_coefs(*o.split(\" \")) for o in open(file, encoding='latin'))\n        \n    return embeddings_index","metadata":{"execution":{"iopub.status.busy":"2022-01-07T00:24:55.985009Z","iopub.execute_input":"2022-01-07T00:24:55.985286Z","iopub.status.idle":"2022-01-07T00:24:55.991155Z","shell.execute_reply.started":"2022-01-07T00:24:55.985241Z","shell.execute_reply":"2022-01-07T00:24:55.990391Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Ta sẽ load thử file nhúng glove Embeddings. Thời gian load file này khá lâu do phải upload hết các chữ thành các vector số để tham chiếu","metadata":{}},{"cell_type":"code","source":"embed_glove = load_embed(glove)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:45:51.781222Z","iopub.execute_input":"2022-01-06T17:45:51.781485Z","iopub.status.idle":"2022-01-06T17:50:12.400069Z","shell.execute_reply.started":"2022-01-06T17:45:51.781450Z","shell.execute_reply":"2022-01-06T17:50:12.399234Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"ví dụ 1 chữ được embed_glove","metadata":{}},{"cell_type":"code","source":"print(embed_glove['the'])","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:50:12.403455Z","iopub.execute_input":"2022-01-06T17:50:12.403679Z","iopub.status.idle":"2022-01-06T17:50:12.414671Z","shell.execute_reply.started":"2022-01-06T17:50:12.403653Z","shell.execute_reply":"2022-01-06T17:50:12.413943Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Nhúng glove embeddings vào các từ được mã hóa thành tokenizer lưu trong word_index. Thay vì lưu vector số tương ứng với các từ, thì ta lưu các vector số hóa tương ứng mã hóa từ ý. Vậy thì file nhúng có đảm bảo mã hóa được tất cả mọi từ có trong câu hỏi không? câu trả lời là không. Vậy làm thế nào để mã hóa từ không có trong file nhúng. Câu trả lời ta lấy phân phối chuẩn của vector số các từ đã biết","metadata":{}},{"cell_type":"code","source":"# Lấy phân phối chuẩn của các từ đã biết ghi vào mọi từ \nall_embs = np.stack(embed_glove.values())\nemb_mean,emb_std = all_embs.mean(), all_embs.std()\nembed_size = all_embs.shape[1]\nword_index = tokenizer.word_index\nnb_words = min(max_features, len(word_index))\nembedding_matrix = np.random.normal(emb_mean, emb_std, (nb_words, embed_size))\n\n# Replace nếu như từ đó được định nghĩa trong file nhúng\nfor word, i in word_index.items():\n    if i >= max_features: continue\n    embedding_vector = embed_glove.get(word)\n    if embedding_vector is not None: embedding_matrix[i] = embedding_vector","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:50:12.415804Z","iopub.execute_input":"2022-01-06T17:50:12.416487Z","iopub.status.idle":"2022-01-06T17:50:22.424883Z","shell.execute_reply.started":"2022-01-06T17:50:12.416449Z","shell.execute_reply":"2022-01-06T17:50:22.424144Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(word_index['the'])\nprint(embed_glove['the'])\nprint(embedding_matrix[2])","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:50:22.453919Z","iopub.execute_input":"2022-01-06T17:50:22.454132Z","iopub.status.idle":"2022-01-06T17:50:22.471777Z","shell.execute_reply.started":"2022-01-06T17:50:22.454107Z","shell.execute_reply":"2022-01-06T17:50:22.470931Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Xây dựng model lại với glove_embeddings. Tương tự như xây dựng model GRU trên","metadata":{}},{"cell_type":"code","source":"inp = Input(shape=(maxlen,))\nx = Embedding(max_features, embed_size, weights=[embedding_matrix])(inp)\n# add weights = [embedding_matrix]\nx = Bidirectional(GRU(64, return_sequences=True))(x)\nx = GlobalMaxPool1D()(x)\nx = Dense(16, activation=\"relu\")(x)\nx = Dropout(0.1)(x)\nx = Dense(1, activation=\"sigmoid\")(x)\nmodel = Model(inputs=inp, outputs=x)\nmodel.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:50:22.473335Z","iopub.execute_input":"2022-01-06T17:50:22.473611Z","iopub.status.idle":"2022-01-06T17:50:22.986029Z","shell.execute_reply.started":"2022-01-06T17:50:22.473575Z","shell.execute_reply":"2022-01-06T17:50:22.985294Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(model.summary())","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:50:22.987419Z","iopub.execute_input":"2022-01-06T17:50:22.987660Z","iopub.status.idle":"2022-01-06T17:50:22.999345Z","shell.execute_reply.started":"2022-01-06T17:50:22.987628Z","shell.execute_reply":"2022-01-06T17:50:22.998530Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"model.fit(train_X, train_y, batch_size=512, epochs=2, validation_data=(val_X, val_y))","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:50:23.000275Z","iopub.execute_input":"2022-01-06T17:50:23.000817Z","iopub.status.idle":"2022-01-06T17:53:47.908471Z","shell.execute_reply.started":"2022-01-06T17:50:23.000781Z","shell.execute_reply":"2022-01-06T17:53:47.907692Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"pred_glove_val_y = model.predict([val_X], batch_size=1024, verbose=1)\nfor thresh in np.arange(0.1, 0.501, 0.01):\n    thresh = np.round(thresh, 2)\n    print(\"F1 score at threshold {0} is {1}\".format(thresh, metrics.f1_score(val_y, (pred_glove_val_y>thresh).astype(int))))","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:53:47.909906Z","iopub.execute_input":"2022-01-06T17:53:47.910158Z","iopub.status.idle":"2022-01-06T17:53:52.292547Z","shell.execute_reply.started":"2022-01-06T17:53:47.910125Z","shell.execute_reply":"2022-01-06T17:53:52.291844Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Kết quả có vẻ tốt hơn so với việc không dùng Embeddings","metadata":{}},{"cell_type":"code","source":"# Lưu lại predict\npred_glove_test_y = model.predict([test_X], batch_size=1024, verbose=1)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:53:52.294710Z","iopub.execute_input":"2022-01-06T17:53:52.295491Z","iopub.status.idle":"2022-01-06T17:54:02.670581Z","shell.execute_reply.started":"2022-01-06T17:53:52.295462Z","shell.execute_reply":"2022-01-06T17:54:02.669816Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"del word_index, embed_glove, all_embs, embedding_matrix, model, inp, x\nimport gc; gc.collect()\ntime.sleep(10)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:54:02.674385Z","iopub.execute_input":"2022-01-06T17:54:02.674588Z","iopub.status.idle":"2022-01-06T17:54:14.006098Z","shell.execute_reply.started":"2022-01-06T17:54:02.674563Z","shell.execute_reply":"2022-01-06T17:54:14.005324Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Tiếp theo ta thử với WikiNews FastText Embedding","metadata":{}},{"cell_type":"markdown","source":"Load wiki news fasttext embeddings","metadata":{}},{"cell_type":"code","source":"embed_wiki_news = load_embed(wiki_news)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:54:14.242936Z","iopub.execute_input":"2022-01-06T17:54:14.243197Z","iopub.status.idle":"2022-01-06T17:56:01.899804Z","shell.execute_reply.started":"2022-01-06T17:54:14.243168Z","shell.execute_reply":"2022-01-06T17:56:01.899057Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"all_embs = np.stack(embed_wiki_news.values())\nemb_mean,emb_std = all_embs.mean(), all_embs.std()\nembed_size = all_embs.shape[1]\n\nword_index = tokenizer.word_index\nnb_words = min(max_features, len(word_index))\nembedding_matrix = np.random.normal(emb_mean, emb_std, (nb_words, embed_size))\n\nfor word, i in word_index.items():\n    if i >= max_features: continue\n    embedding_vector = embed_wiki_news.get(word)\n    if embedding_vector is not None: embedding_matrix[i] = embedding_vector","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:56:01.903034Z","iopub.execute_input":"2022-01-06T17:56:01.903238Z","iopub.status.idle":"2022-01-06T17:56:07.307266Z","shell.execute_reply.started":"2022-01-06T17:56:01.903213Z","shell.execute_reply":"2022-01-06T17:56:07.306509Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"inp = Input(shape=(maxlen,))\nx = Embedding(max_features, embed_size, weights=[embedding_matrix])(inp)\n# add weights = [embedding_matrix]\nx = Bidirectional(GRU(64, return_sequences=True))(x)\nx = GlobalMaxPool1D()(x)\nx = Dense(16, activation=\"relu\")(x)\nx = Dropout(0.1)(x)\nx = Dense(1, activation=\"sigmoid\")(x)\nmodel = Model(inputs=inp, outputs=x)\nmodel.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:56:07.308545Z","iopub.execute_input":"2022-01-06T17:56:07.308798Z","iopub.status.idle":"2022-01-06T17:56:07.803964Z","shell.execute_reply.started":"2022-01-06T17:56:07.308766Z","shell.execute_reply":"2022-01-06T17:56:07.803207Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"model.fit(train_X, train_y, batch_size=512, epochs=2, validation_data=(val_X, val_y))","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:56:07.805104Z","iopub.execute_input":"2022-01-06T17:56:07.805346Z","iopub.status.idle":"2022-01-06T17:59:01.330416Z","shell.execute_reply.started":"2022-01-06T17:56:07.805315Z","shell.execute_reply":"2022-01-06T17:59:01.329709Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"pred_wiki_news_val_y = model.predict([val_X], batch_size=1024, verbose=1)\nfor thresh in np.arange(0.1, 0.501, 0.01):\n    thresh = np.round(thresh, 2)\n    print(\"F1 score at threshold {0} is {1}\".format(thresh, metrics.f1_score(val_y, (pred_glove_val_y>thresh).astype(int))))","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:59:01.331780Z","iopub.execute_input":"2022-01-06T17:59:01.332090Z","iopub.status.idle":"2022-01-06T17:59:05.978377Z","shell.execute_reply.started":"2022-01-06T17:59:01.332055Z","shell.execute_reply":"2022-01-06T17:59:05.977556Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Kết quả thu được vẫn tốt hơn so với việc không dùng nhúng, xấp xỉ so với việc dùng Glove Embeddings","metadata":{}},{"cell_type":"code","source":"# Lưu lại predict\npred_wiki_news_test_y = model.predict([test_X], batch_size=1024, verbose=1)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:59:05.979769Z","iopub.execute_input":"2022-01-06T17:59:05.980059Z","iopub.status.idle":"2022-01-06T17:59:16.364022Z","shell.execute_reply.started":"2022-01-06T17:59:05.980021Z","shell.execute_reply":"2022-01-06T17:59:16.363247Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"del embed_wiki_news, all_embs, embedding_matrix, model, inp, x\nimport gc; gc.collect()\ntime.sleep(10)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:59:16.365653Z","iopub.execute_input":"2022-01-06T17:59:16.365912Z","iopub.status.idle":"2022-01-06T17:59:27.219673Z","shell.execute_reply.started":"2022-01-06T17:59:16.365857Z","shell.execute_reply":"2022-01-06T17:59:27.218909Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Thử với Paragram Embeddings","metadata":{}},{"cell_type":"code","source":"embed_paragram = load_embed(paragram)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T17:59:27.254649Z","iopub.execute_input":"2022-01-06T17:59:27.254898Z","iopub.status.idle":"2022-01-06T18:02:33.573432Z","shell.execute_reply.started":"2022-01-06T17:59:27.254857Z","shell.execute_reply":"2022-01-06T18:02:33.572388Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"all_embs = np.stack(embed_paragram.values())\nemb_mean,emb_std = all_embs.mean(), all_embs.std()\nembed_size = all_embs.shape[1]\n\nword_index = tokenizer.word_index\nnb_words = min(max_features, len(word_index))\nembedding_matrix = np.random.normal(emb_mean, emb_std, (nb_words, embed_size))\n\nfor word, i in word_index.items():\n    if i >= max_features: continue\n    embedding_vector = embed_paragram.get(word)\n    if embedding_vector is not None: embedding_matrix[i] = embedding_vector","metadata":{"execution":{"iopub.status.busy":"2022-01-06T18:02:33.574855Z","iopub.execute_input":"2022-01-06T18:02:33.575132Z","iopub.status.idle":"2022-01-06T18:02:41.860686Z","shell.execute_reply.started":"2022-01-06T18:02:33.575096Z","shell.execute_reply":"2022-01-06T18:02:41.859917Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"inp = Input(shape=(maxlen,))\nx = Embedding(max_features, embed_size, weights=[embedding_matrix])(inp)\n# add weights = [embedding_matrix]\nx = Bidirectional(GRU(64, return_sequences=True))(x)\nx = GlobalMaxPool1D()(x)\nx = Dense(16, activation=\"relu\")(x)\nx = Dropout(0.1)(x)\nx = Dense(1, activation=\"sigmoid\")(x)\nmodel = Model(inputs=inp, outputs=x)\nmodel.compile(loss='binary_crossentropy', optimizer='adam', metrics=['accuracy'])","metadata":{"execution":{"iopub.status.busy":"2022-01-06T18:02:41.886674Z","iopub.execute_input":"2022-01-06T18:02:41.888693Z","iopub.status.idle":"2022-01-06T18:02:42.403751Z","shell.execute_reply.started":"2022-01-06T18:02:41.888649Z","shell.execute_reply":"2022-01-06T18:02:42.403042Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"model.fit(train_X, train_y, batch_size=512, epochs=2, validation_data=(val_X, val_y))","metadata":{"execution":{"iopub.status.busy":"2022-01-06T18:02:42.404970Z","iopub.execute_input":"2022-01-06T18:02:42.405209Z","iopub.status.idle":"2022-01-06T18:06:07.438440Z","shell.execute_reply.started":"2022-01-06T18:02:42.405177Z","shell.execute_reply":"2022-01-06T18:06:07.437709Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"pred_paragram_val_y = model.predict([val_X], batch_size=1024, verbose=1)\nfor thresh in np.arange(0.1, 0.501, 0.01):\n    thresh = np.round(thresh, 2)\n    print(\"F1 score at threshold {0} is {1}\".format(thresh, metrics.f1_score(val_y, (pred_glove_val_y>thresh).astype(int))))","metadata":{"execution":{"iopub.status.busy":"2022-01-06T18:06:07.439636Z","iopub.execute_input":"2022-01-06T18:06:07.441157Z","iopub.status.idle":"2022-01-06T18:06:11.167944Z","shell.execute_reply.started":"2022-01-06T18:06:07.441114Z","shell.execute_reply":"2022-01-06T18:06:11.167161Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"pred_paragram_test_y = model.predict([test_X], batch_size=1024, verbose=1)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T18:06:11.169341Z","iopub.execute_input":"2022-01-06T18:06:11.169605Z","iopub.status.idle":"2022-01-06T18:06:16.508763Z","shell.execute_reply.started":"2022-01-06T18:06:11.169568Z","shell.execute_reply":"2022-01-06T18:06:16.508058Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Kết quả vẫn xấp xỉ với 2 tool nhúng trước. Ta thử kết hợp cả 3 tool nhúng để xem kết quả có cải thiện hay không","metadata":{}},{"cell_type":"code","source":"pred_val_y = 0.3*pred_glove_val_y + 0.3*pred_wiki_news_val_y + 0.34*pred_paragram_val_y \nfor thresh in np.arange(0.1, 0.501, 0.01):\n    thresh = np.round(thresh, 2)\n    print(\"F1 score at threshold {0} is {1}\".format(thresh, metrics.f1_score(val_y, (pred_val_y>thresh).astype(int))))","metadata":{"execution":{"iopub.status.busy":"2022-01-06T18:06:16.510188Z","iopub.execute_input":"2022-01-06T18:06:16.510380Z","iopub.status.idle":"2022-01-06T18:06:17.876199Z","shell.execute_reply.started":"2022-01-06T18:06:16.510356Z","shell.execute_reply":"2022-01-06T18:06:17.875456Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"pred_paragram_test_y = model.predict([test_X], batch_size=1024, verbose=1)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T18:06:17.877478Z","iopub.execute_input":"2022-01-06T18:06:17.877883Z","iopub.status.idle":"2022-01-06T18:06:28.255176Z","shell.execute_reply.started":"2022-01-06T18:06:17.877828Z","shell.execute_reply":"2022-01-06T18:06:28.254146Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Tốt hơn một chút. Ta quyết định lấy kết quả chính là sự kết hợp của 3 model này. Chọn threshold là 0.3 vì nó cho F1-score là tốt nhất.","metadata":{}},{"cell_type":"code","source":"pred_test_y = 0.33*pred_glove_test_y + 0.33*pred_wiki_news_test_y + 0.34*pred_paragram_test_y\npred_test_y = (pred_test_y>0.3).astype(int)\nout_df = pd.DataFrame({\"qid\":test_df[\"qid\"].values})\nout_df['prediction'] = pred_test_y\nout_df.to_csv(\"submission.csv\", index=False)","metadata":{"execution":{"iopub.status.busy":"2022-01-06T18:06:28.256773Z","iopub.execute_input":"2022-01-06T18:06:28.257045Z","iopub.status.idle":"2022-01-06T18:06:29.069275Z","shell.execute_reply.started":"2022-01-06T18:06:28.257008Z","shell.execute_reply":"2022-01-06T18:06:29.068512Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"out_df","metadata":{"execution":{"iopub.status.busy":"2022-01-06T18:06:29.072394Z","iopub.execute_input":"2022-01-06T18:06:29.072608Z","iopub.status.idle":"2022-01-06T18:06:29.086661Z","shell.execute_reply.started":"2022-01-06T18:06:29.072582Z","shell.execute_reply":"2022-01-06T18:06:29.085919Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Test 10 câu hỏi không chân thành","metadata":{}},{"cell_type":"code","source":"dem = 0;\nfor id,i in enumerate(pred_test_y):\n    if i == 1:\n        print(test_df['question_text'][id])\n        dem = dem + 1\n        if dem == 10:\n            break","metadata":{"execution":{"iopub.status.busy":"2022-01-06T18:06:29.089607Z","iopub.execute_input":"2022-01-06T18:06:29.090155Z","iopub.status.idle":"2022-01-06T18:06:29.098255Z","shell.execute_reply.started":"2022-01-06T18:06:29.090127Z","shell.execute_reply":"2022-01-06T18:06:29.097268Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Nhận xét: Mô hình hoạt động đúng như kì vọng","metadata":{}},{"cell_type":"code","source":"# Tổng hợp tất cả các câu hỏi lại để lấy ra tất cả các từ vựng trong đó\ndf = pd.concat([test_df, train_df])","metadata":{"execution":{"iopub.status.busy":"2022-01-07T00:24:12.086351Z","iopub.execute_input":"2022-01-07T00:24:12.086988Z","iopub.status.idle":"2022-01-07T00:24:12.177991Z","shell.execute_reply.started":"2022-01-07T00:24:12.086950Z","shell.execute_reply":"2022-01-07T00:24:12.177280Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# load lại các file nhúng\nembed_glove = load_embed(glove)\nembed_wiki_news = load_embed(wiki_news)\nembed_paragram = load_embed(paragram)","metadata":{"execution":{"iopub.status.busy":"2022-01-07T00:40:49.631914Z","iopub.execute_input":"2022-01-07T00:40:49.632783Z","iopub.status.idle":"2022-01-07T00:50:06.277190Z","shell.execute_reply.started":"2022-01-07T00:40:49.632738Z","shell.execute_reply":"2022-01-07T00:50:06.276279Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Xây dựng vocab chứa mọi từ vựng trong mọi câu hỏi và số lần xuất hiện của chúng\ndef build_vocab(texts):\n    sentences = texts.apply(lambda x: x.split()).values\n    vocab = {}\n    for sentence in sentences:\n        for word in sentence:\n            try:\n                vocab[word] += 1\n            except KeyError:\n                vocab[word] = 1\n    return vocab","metadata":{"execution":{"iopub.status.busy":"2022-01-07T00:39:03.913049Z","iopub.execute_input":"2022-01-07T00:39:03.913318Z","iopub.status.idle":"2022-01-07T00:39:03.921181Z","shell.execute_reply.started":"2022-01-07T00:39:03.913284Z","shell.execute_reply":"2022-01-07T00:39:03.920357Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"vocab = build_vocab(df['question_text'])","metadata":{"execution":{"iopub.status.busy":"2022-01-07T00:39:03.922501Z","iopub.execute_input":"2022-01-07T00:39:03.922818Z","iopub.status.idle":"2022-01-07T00:39:14.905106Z","shell.execute_reply.started":"2022-01-07T00:39:03.922779Z","shell.execute_reply":"2022-01-07T00:39:14.904378Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import operator \ndef check_coverage(vocab, embeddings_index):\n    known_words = {}\n    unknown_words = {}\n    nb_known_words = 0\n    nb_unknown_words = 0\n    for word in vocab.keys():\n        try:\n            known_words[word] = embeddings_index[word]\n            nb_known_words += vocab[word]\n        except:\n            unknown_words[word] = vocab[word]\n            nb_unknown_words += vocab[word]\n            pass\n\n    print('Found embeddings for {:.2%} of vocab'.format(len(known_words) / len(vocab)))\n    print('Found embeddings for  {:.2%} of all text'.format(nb_known_words / (nb_known_words + nb_unknown_words)))\n    unknown_words = sorted(unknown_words.items(), key=operator.itemgetter(1))[::-1]\n\n    return unknown_words","metadata":{"execution":{"iopub.status.busy":"2022-01-07T00:50:06.385184Z","iopub.execute_input":"2022-01-07T00:50:06.385640Z","iopub.status.idle":"2022-01-07T00:50:06.397239Z","shell.execute_reply.started":"2022-01-07T00:50:06.385592Z","shell.execute_reply":"2022-01-07T00:50:06.396499Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(\"Glove : \")\noov_glove = check_coverage(vocab, embed_glove)\nprint(\"Paragram : \")\noov_paragram = check_coverage(vocab, embed_paragram)\nprint(\"Wiki news FastText : \")\noov_fasttext = check_coverage(vocab, embed_wiki_news)","metadata":{"execution":{"iopub.status.busy":"2022-01-07T00:39:14.916357Z","iopub.execute_input":"2022-01-07T00:39:14.916636Z","iopub.status.idle":"2022-01-07T00:39:16.987250Z","shell.execute_reply.started":"2022-01-07T00:39:14.916599Z","shell.execute_reply":"2022-01-07T00:39:16.986434Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Nhận xét\nKết quả cho ta thấy:\nViệc sử dụng nhúng đem lại kết quả cao hơn khi không sử dụng\n3 loại nhúng về cơ bản cho hiệu suất là như nhau\n\nĐặt vắn đề: Với kết quả như vậy liệu ta có thể cải tiến được nó cho đạt hiệu suất cao hơn được hay không? cùng quay lại khi xây dựng ở tầng embedding ta có khi từ đó không xuất hiện trong file nhúng ta sử dụng phân phối chuẩn để embedding chúng. Mặc dù chúng ta đã sửa hết các lỗi chính ta hay loại bỏ dư thừa như link thì file nhúng của chúng ta thực sự đã chứa bao nhiêu % các từ đã biết trên tổng số các câu hỏi? Sau khi kiểm tra qua về độ phủ của chúng, ta thấy độ phủ lần lượt qua các file glove-paragram-wiki lần lượt là 88,16%,72,21%,87,66%. Chưa phủ được toàn bộ word trong file train và test điều này cũng dễ hiểu vì trong tập train và test đó là dữ liệu người dùng nên có lỗi chính tả hay từ địa phương cũng là dễ hiểu. Để tăng hiệu suất của mô hình ta nên cố gắng phủ được càng cao số từ trong tập train và test càng tốt-> bằng cách này em nghĩ ta sẽ thu được mô hình tốt hơn","metadata":{}},{"cell_type":"markdown","source":"# Kết quả\n\n![Untitled.png](attachment:0509b331-49dd-47f1-81a7-c7ce058807ef.png)","metadata":{},"attachments":{"0509b331-49dd-47f1-81a7-c7ce058807ef.png":{"image/png":"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"}}}]}