{"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":"code","source":"import numpy as np\nimport pandas as pd \nimport matplotlib.pyplot as plt\n\nimport re\nimport nltk\nfrom nltk.tokenize import word_tokenize\nfrom nltk.corpus import stopwords\n\nfrom sklearn.feature_extraction.text import CountVectorizer\nfrom sklearn.linear_model import LogisticRegression\nfrom sklearn.metrics import accuracy_score, f1_score,recall_score, classification_report, confusion_matrix\nfrom sklearn.model_selection import train_test_split","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","execution":{"iopub.status.busy":"2022-01-04T16:36:13.706389Z","iopub.execute_input":"2022-01-04T16:36:13.706757Z","iopub.status.idle":"2022-01-04T16:36:14.958377Z","shell.execute_reply.started":"2022-01-04T16:36:13.706675Z","shell.execute_reply":"2022-01-04T16:36:14.957648Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Mô tả bài toán\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. Một thách thức lớn là loại bỏ những câu hỏi thiếu chân thành - những câu hỏi được đặt ra nhằm mục địch gây tranh cãi, nội dung độc hại hoặc có ý định đưa ra một tuyên bố hơn là tìm kiếm những câu trả lời hữu ích. Quora muốn giải quyết vấn đề này để giữ cho nền tảng của họ trở thành một nơi mà người dùng có thể cảm thấy an toàn khi chia sẻ kiến thức.\n* Mục tiêu: Phân loại câu hỏi Sincere và Insincere trên Quora\n* Input: Câu hỏi dạng text\n* Output: 0/1 (Sincere/ Insincere)","metadata":{}},{"cell_type":"markdown","source":"# Phân tích dữ liệu","metadata":{}},{"cell_type":"code","source":"train_data = pd.read_csv('/kaggle/input/quora-insincere-questions-classification/train.csv')\ntest_data = pd.read_csv('/kaggle/input/quora-insincere-questions-classification/test.csv')","metadata":{"execution":{"iopub.status.busy":"2022-01-04T16:36:14.960135Z","iopub.execute_input":"2022-01-04T16:36:14.960383Z","iopub.status.idle":"2022-01-04T16:36:20.352249Z","shell.execute_reply.started":"2022-01-04T16:36:14.960351Z","shell.execute_reply":"2022-01-04T16:36:20.351502Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_data","metadata":{"execution":{"iopub.status.busy":"2022-01-04T16:36:20.353546Z","iopub.execute_input":"2022-01-04T16:36:20.353781Z","iopub.status.idle":"2022-01-04T16:36:20.376183Z","shell.execute_reply.started":"2022-01-04T16:36:20.353748Z","shell.execute_reply":"2022-01-04T16:36:20.375549Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_data.info()","metadata":{"execution":{"iopub.status.busy":"2022-01-04T16:36:20.377438Z","iopub.execute_input":"2022-01-04T16:36:20.377692Z","iopub.status.idle":"2022-01-04T16:36:20.646748Z","shell.execute_reply.started":"2022-01-04T16:36:20.377659Z","shell.execute_reply":"2022-01-04T16:36:20.64597Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"test_data","metadata":{"execution":{"iopub.status.busy":"2022-01-04T16:36:20.64897Z","iopub.execute_input":"2022-01-04T16:36:20.649652Z","iopub.status.idle":"2022-01-04T16:36:20.65972Z","shell.execute_reply.started":"2022-01-04T16:36:20.649614Z","shell.execute_reply":"2022-01-04T16:36:20.65899Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"test_data.info()","metadata":{"execution":{"iopub.status.busy":"2022-01-04T16:36:20.661181Z","iopub.execute_input":"2022-01-04T16:36:20.661656Z","iopub.status.idle":"2022-01-04T16:36:20.744319Z","shell.execute_reply.started":"2022-01-04T16:36:20.661616Z","shell.execute_reply":"2022-01-04T16:36:20.743565Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"**Nhận xét:** Tập dữ liệu không có giá trị null\n\n**File train** gồm: \n* 1306122 dòng x 3 cột (qid, question_text, target)\n\n**File test** gồm: \n* 375806 dòng x 2 cột (qid, question_text)\n\n**Chú thích**\n* qid: id của câu hỏi\n* question_text: nội dung câu hỏi\n* target: phân loại câu hỏi, target = 0 là câu hỏi Sincere, target = 1 là câu hỏi Insincere","metadata":{}},{"cell_type":"code","source":"train_data['target'].value_counts()","metadata":{},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"**Nhận xét**\n* Có 1,225,312 câu hỏi target bằng 0, tương ứng là các câu hỏi Sincere.\n* Có 80,810 câu hỏi target bằng 1, tương ứng là các câu hỏi Insincere.","metadata":{}},{"cell_type":"code","source":"# Pie chart biểu diễn tỉ lệ câu hỏi\ntoxic_question = train_data[train_data.target == 1]\nnon_toxic_question = train_data[train_data.target == 0]\n\nlabel = 'Toxic', 'Non-toxic'\nsize = [(toxic_question.shape[0] / train_data.shape[0]) * 100, (non_toxic_question.shape[0] / train_data.shape[0]) * 100]\nplt.pie(size,labels = label,colors=[\"m\", \"c\"], autopct=\"%.2f%%\")\nplt.axis('equal')\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2022-01-04T16:36:20.745616Z","iopub.execute_input":"2022-01-04T16:36:20.745843Z","iopub.status.idle":"2022-01-04T16:36:20.941062Z","shell.execute_reply.started":"2022-01-04T16:36:20.745807Z","shell.execute_reply":"2022-01-04T16:36:20.940282Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"**Nhận xét**\n* Dựa vào biểu đồ trên ta có thể thấy rằng có **6.19%** câu hỏi được đánh **nhãn 1** và **93.81%** câu hỏi đánh **nhãn 0**.\n* Dữ liệu đang mất cân bằng, với tỉ lệ này, accuracy có thể rất cao mà không cần dùng tới mô hình (khi đoán toàn bộ nhãn là 0, accuracy lên tới gần 94%).\\\n=> Chọn **F1-score** làm chỉ số đánh giá mô hình.\n\n![](https://miro.medium.com/max/954/1*2dfEH04P9wffwMYjKP5hUg.png)","metadata":{}},{"cell_type":"markdown","source":"# Xử lý mất cân bằng dữ liệu\nDùng Under Sampling để giảm số lượng câu Sincere giúp dữ liệu cân bằng hơn\n* Ưu điểm: làm cân bằng mẫu một cách nhanh chóng, dễ dàng tiến hành thực hiện mà không cần đến thuật toán giả lập mẫu.\n* Nhược điểm: kích thước mẫu sẽ bị giảm đáng kể. ","metadata":{}},{"cell_type":"code","source":"#train_data_US = pd.concat([resample(non_toxic_question, replace = True, n_samples = len(toxic_question)*4), toxic_question])\n#train_data_US\n\n# Sau khi giảm số lượng câu xuống tỉ lệ 4 câu Sincere:1 câu Insincere thì score thấp hơn do mất nhiều dữ liệu khiến mô hình dự đoán thiếu chính xác.","metadata":{"execution":{"iopub.status.busy":"2022-01-04T16:36:20.942803Z","iopub.execute_input":"2022-01-04T16:36:20.943226Z","iopub.status.idle":"2022-01-04T16:36:20.947778Z","shell.execute_reply.started":"2022-01-04T16:36:20.943188Z","shell.execute_reply":"2022-01-04T16:36:20.946946Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"y = train_data['target']\ny.value_counts().plot(kind='bar', rot=0)","metadata":{"execution":{"iopub.status.busy":"2022-01-04T16:36:20.949252Z","iopub.execute_input":"2022-01-04T16:36:20.950043Z","iopub.status.idle":"2022-01-04T16:36:21.137306Z","shell.execute_reply.started":"2022-01-04T16:36:20.950008Z","shell.execute_reply":"2022-01-04T16:36:21.136675Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Tiền xử lý\n* Chuyển về dạng viết thường\n* Xóa ký hiệu đặc biệt\n* Tách các từ\n* Loại bỏ stopwords\n* Chia tập dữ liệu\n* Text Vectorization","metadata":{}},{"cell_type":"code","source":"nltk.download('stopwords', 'punkt')\nsw = stopwords.words('english')\nsw.remove('not')\nprint(sw)","metadata":{"execution":{"iopub.status.busy":"2022-01-04T16:36:21.138456Z","iopub.execute_input":"2022-01-04T16:36:21.138705Z","iopub.status.idle":"2022-01-04T16:36:41.180613Z","shell.execute_reply.started":"2022-01-04T16:36:21.138671Z","shell.execute_reply":"2022-01-04T16:36:41.179713Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def normalize_txt(text):\n    # chuyển về dạng viết thường, xóa khoảng trắng (whitespaces)\n    text = text.strip().lower()\n\n    # xóa số\n    text = re.sub(r'\\d+', '', text)\n    \n    # xóa tag HTML    \n    text = re.sub(re.compile('<.*?>'), '', text)\n    \n    # xóa ký tự đặc biệt\n    text = re.sub(r'[^a-zA-Z\\s]', '', text, re.I|re.A)\n    \n    # tokenize dữ liệu\n    tokens = word_tokenize(text)\n    \n    # bỏ stopwords\n    tokens_noSW = [w for w in tokens if not w in sw]\n    text = ' '.join(tokens_noSW)\n\n    return text","metadata":{"execution":{"iopub.status.busy":"2022-01-04T16:36:41.181701Z","iopub.execute_input":"2022-01-04T16:36:41.183077Z","iopub.status.idle":"2022-01-04T16:36:41.18945Z","shell.execute_reply.started":"2022-01-04T16:36:41.183028Z","shell.execute_reply":"2022-01-04T16:36:41.188459Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_data['normalize'] = train_data['question_text'].apply(normalize_txt)\ntrain_data","metadata":{"execution":{"iopub.status.busy":"2022-01-04T16:36:41.191166Z","iopub.execute_input":"2022-01-04T16:36:41.19141Z","iopub.status.idle":"2022-01-04T16:40:52.400836Z","shell.execute_reply.started":"2022-01-04T16:36:41.191377Z","shell.execute_reply":"2022-01-04T16:40:52.399099Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"X = train_data['normalize']\n\n# Chia tập train theo tỉ lệ: 75% train - 25% test\nX_train, X_test, y_train, y_test = train_test_split(X, y, test_size=0.25, random_state=0)\n\nprint('X_train: ', X_train.shape)\nprint('X_test: ', X_test.shape)","metadata":{"execution":{"iopub.status.busy":"2022-01-04T16:40:52.402119Z","iopub.execute_input":"2022-01-04T16:40:52.402449Z","iopub.status.idle":"2022-01-04T16:40:52.710184Z","shell.execute_reply.started":"2022-01-04T16:40:52.402413Z","shell.execute_reply":"2022-01-04T16:40:52.70944Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"**Vector hóa dữ liệu:** \nĐể sử dụng dữ liệu văn bản cho mô hình dự đoán, văn bản cần được mã hóa dưới dạng số nguyên hoặc giá trị dấu phẩy động, quá trình này được gọi là trích xuất đặc trưng (hoặc vectơ hóa). Hiện này có hai cách thực hiện kỹ thuật này là:\n* Tính toán số lần xuất hiện của từ trong văn bản\n* Tính toán tần suất xuất hiện của từ trong văn bản\n\n**CountVectorizer**: được sử dụng để chia câu hỏi thành các từ, chuyển nó thành một vectơ trên cơ sở tần suất (số lượng) của mỗi từ xuất hiện. Nó cũng cho phép xử lý trước dữ liệu văn bản trước khi tạo biểu diễn vectơ.","metadata":{}},{"cell_type":"code","source":"vectorizer = CountVectorizer(analyzer='word', ngram_range=(1,3))\ncv_train = vectorizer.fit_transform(X_train)\nprint(cv_train.shape)\ncv_test = vectorizer.transform(X_test)\nprint(cv_test.shape)","metadata":{"execution":{"iopub.status.busy":"2022-01-04T16:40:52.712958Z","iopub.execute_input":"2022-01-04T16:40:52.713376Z","iopub.status.idle":"2022-01-04T16:41:06.979812Z","shell.execute_reply.started":"2022-01-04T16:40:52.713337Z","shell.execute_reply":"2022-01-04T16:41:06.978206Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Xây dựng mô hình\n**Logistic Regression**\n* Mô hình đơn giản, phổ biến cho bài toán phân lớp\n* Hồi quy logistic là một phương pháp phân tích thống kê được sử dụng để dự đoán giá trị dữ liệu dựa trên các quan sát trước đó của tập dữ liệu\\\n=> Có thể thấy là với bài toán phân loại nhị phân như bài toán này, việc sử dụng **Logistic Regression** là khá phù hợp.\n* Đầu ra dự đoán của mô hình Logistic Regression: $f(\\mathbf{x})=\\mathbf{w}^{T} \\mathbf{x}$\n* Trong số các hàm số logistic thì hàm sigmoid được sử dụng nhiều nhất, vì nó bị chặn trong khoảng (0,1): $f(s)=\\frac{1}{1+e^{-s}} \\triangleq \\sigma(s)$","metadata":{}},{"cell_type":"code","source":"LR = LogisticRegression(solver='liblinear', class_weight=\"balanced\")\nLR.fit(cv_train, y_train)\n\ntest_pred = LR.predict(cv_test)\nprint(classification_report(y_test, test_pred))","metadata":{"execution":{"iopub.status.busy":"2022-01-04T16:41:06.981052Z","iopub.execute_input":"2022-01-04T16:41:06.981339Z","iopub.status.idle":"2022-01-04T16:41:50.325446Z","shell.execute_reply.started":"2022-01-04T16:41:06.981296Z","shell.execute_reply":"2022-01-04T16:41:50.324695Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"test_data['normalize'] = test_data['question_text'].apply(normalize_txt)\n\nx = vectorizer.transform(test_data['normalize'])\npred = LR.predict(x)\n\ntest_data['prediction'] = pred\ntest_data","metadata":{"execution":{"iopub.status.busy":"2022-01-04T16:41:50.326553Z","iopub.execute_input":"2022-01-04T16:41:50.326989Z","iopub.status.idle":"2022-01-04T16:43:06.36242Z","shell.execute_reply.started":"2022-01-04T16:41:50.326949Z","shell.execute_reply":"2022-01-04T16:43:06.361726Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"result = test_data[['qid', 'prediction']]\nresult.to_csv('submission.csv', index=False)\nresult","metadata":{"execution":{"iopub.status.busy":"2022-01-04T16:43:06.363832Z","iopub.execute_input":"2022-01-04T16:43:06.364315Z","iopub.status.idle":"2022-01-04T16:43:07.181401Z","shell.execute_reply.started":"2022-01-04T16:43:06.364279Z","shell.execute_reply":"2022-01-04T16:43:07.180724Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Báo cáo kết quả\n* Private score: 0.62087\n* Public score: 0.60972\n![image.png](attachment:71ec642f-9133-4524-92ee-82d3a8933f93.png)","metadata":{},"attachments":{"71ec642f-9133-4524-92ee-82d3a8933f93.png":{"image/png":"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"}}}]}