{"metadata":{"kaggle":{"accelerator":"gpu","dataSources":[{"sourceId":50160,"databundleVersionId":7921029,"sourceType":"competition"},{"sourceId":8023315,"sourceType":"datasetVersion","datasetId":4728096}],"dockerImageVersionId":30665,"isInternetEnabled":false,"language":"python","sourceType":"notebook","isGpuEnabled":true},"kernelspec":{"display_name":"Python 3","language":"python","name":"python3"},"language_info":{"name":"python","version":"3.10.13","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"papermill":{"default_parameters":{},"duration":2163.521994,"end_time":"2024-04-02T06:25:50.949986","environment_variables":{},"exception":null,"input_path":"__notebook__.ipynb","output_path":"__notebook__.ipynb","parameters":{},"start_time":"2024-04-02T05:49:47.427992","version":"2.5.0"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"!pip install /kaggle/input/polars-0-20-18/polars-0.20.18-cp38-abi3-manylinux_2_17_x86_64.manylinux2014_x86_64.whl\n","metadata":{"papermill":{"duration":0.019685,"end_time":"2024-04-02T05:49:50.271384","exception":false,"start_time":"2024-04-02T05:49:50.251699","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-04-08T11:33:20.815067Z","iopub.execute_input":"2024-04-08T11:33:20.815310Z","iopub.status.idle":"2024-04-08T11:34:05.413474Z","shell.execute_reply.started":"2024-04-08T11:33:20.815287Z","shell.execute_reply":"2024-04-08T11:34:05.412359Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Adding Target Encoding features for String, Categorical features\n \n
\n 📌   My idea here was to add Target Encoding of data(pl.String, pl.Boolean, pl.Categorical)\n
\n\n
\n 📌   Next Idea : Embedding String data with using BERT\n
\n","metadata":{"papermill":{"duration":0.010745,"end_time":"2024-04-02T05:49:50.293412","exception":false,"start_time":"2024-04-02T05:49:50.282667","status":"completed"},"tags":[]}},{"cell_type":"code","source":"import sys\nfrom pathlib import Path\nimport subprocess\nimport os\nimport gc\nfrom glob import glob\n\nimport numpy as np\nimport pandas as pd\nimport polars as pl\nfrom datetime import datetime\nimport seaborn as sns\nimport matplotlib.pyplot as plt\nfrom itertools import combinations, permutations\n\nimport warnings\nwarnings.filterwarnings('ignore')\n\nROOT = '/kaggle/input/home-credit-credit-risk-model-stability'","metadata":{"_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","papermill":{"duration":2.226914,"end_time":"2024-04-02T05:49:52.531606","exception":false,"start_time":"2024-04-02T05:49:50.304692","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-04-08T11:34:05.415758Z","iopub.execute_input":"2024-04-08T11:34:05.416491Z","iopub.status.idle":"2024-04-08T11:34:06.516185Z","shell.execute_reply.started":"2024-04-08T11:34:05.416450Z","shell.execute_reply":"2024-04-08T11:34:06.515319Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from sklearn.model_selection import TimeSeriesSplit, GroupKFold, StratifiedGroupKFold\nfrom sklearn.base import BaseEstimator, RegressorMixin\nfrom sklearn.metrics import roc_auc_score\nimport lightgbm as lgb\n\nfrom sklearn.preprocessing import OrdinalEncoder\nfrom sklearn.impute import KNNImputer","metadata":{"papermill":{"duration":4.398776,"end_time":"2024-04-02T05:49:56.941597","exception":false,"start_time":"2024-04-02T05:49:52.542821","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-04-08T11:34:06.517351Z","iopub.execute_input":"2024-04-08T11:34:06.517826Z","iopub.status.idle":"2024-04-08T11:34:09.953902Z","shell.execute_reply.started":"2024-04-08T11:34:06.517794Z","shell.execute_reply":"2024-04-08T11:34:09.953105Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"class Pipeline:\n\n def set_table_dtypes(df):\n for col in df.columns:\n if col in [\"case_id\", \"WEEK_NUM\", \"num_group1\", \"num_group2\"]:\n df = df.with_columns(pl.col(col).cast(pl.Int64))\n elif col in [\"date_decision\"]:\n df = df.with_columns(pl.col(col).cast(pl.Date))\n elif col[-1] in (\"P\", \"A\"):\n df = df.with_columns(pl.col(col).cast(pl.Float64))\n elif col[-1] in (\"M\",):\n df = df.with_columns(pl.col(col).cast(pl.String))\n elif col[-1] in (\"D\",):\n df = df.with_columns(pl.col(col).cast(pl.Date))\n return df\n\n def handle_dates(df):\n for col in df.columns:\n if col[-1] in (\"D\",):\n df = df.with_columns(pl.col(col) - pl.col(\"date_decision\")) #!!?\n df = df.with_columns(pl.col(col).dt.total_days()) # t - t-1\n df = df.drop(\"date_decision\", \"MONTH\")\n return df\n\n def filter_cols(df, isnull_ratio):\n for col in df.columns:\n if col not in [\"target\", \"case_id\", \"WEEK_NUM\"]:\n isnull = df[col].is_null().mean()\n if isnull > isnull_ratio:\n df = df.drop(col)\n \n for col in df.columns:\n if (col not in [\"target\", \"case_id\", \"WEEK_NUM\"]) & (df[col].dtype == pl.String):\n freq = df[col].n_unique()\n if (freq == 1) | (freq > 200):\n df = df.drop(col)\n \n return df","metadata":{"papermill":{"duration":0.026262,"end_time":"2024-04-02T05:49:56.979185","exception":false,"start_time":"2024-04-02T05:49:56.952923","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-04-08T11:34:09.956107Z","iopub.execute_input":"2024-04-08T11:34:09.956421Z","iopub.status.idle":"2024-04-08T11:34:09.968362Z","shell.execute_reply.started":"2024-04-08T11:34:09.956396Z","shell.execute_reply":"2024-04-08T11:34:09.967413Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"class Aggregator:\n #Please add or subtract features yourself, be aware that too many features will take up too much space.\n def num_expr(df):\n cols = [col for col in df.columns if col[-1] in (\"P\", \"A\")]\n expr_max = [pl.max(col).alias(f\"max_{col}\") for col in cols]\n \n expr_last = [pl.last(col).alias(f\"last_{col}\") for col in cols]\n #expr_first = [pl.first(col).alias(f\"first_{col}\") for col in cols]\n expr_mean = [pl.mean(col).alias(f\"mean_{col}\") for col in cols]\n\n return expr_max +expr_last+expr_mean\n \n def date_expr(df):\n cols = [col for col in df.columns if col[-1] in (\"D\")]\n expr_max = [pl.max(col).alias(f\"max_{col}\") for col in cols]\n #expr_min = [pl.min(col).alias(f\"min_{col}\") for col in cols]\n expr_last = [pl.last(col).alias(f\"last_{col}\") for col in cols]\n #expr_first = [pl.first(col).alias(f\"first_{col}\") for col in cols]\n expr_mean = [pl.mean(col).alias(f\"mean_{col}\") for col in cols]\n return expr_max +expr_last+expr_mean\n \n def str_expr(df):\n cols = [col for col in df.columns if col[-1] in (\"M\",)]\n expr_max = [pl.max(col).alias(f\"max_{col}\") for col in cols]\n #expr_min = [pl.min(col).alias(f\"min_{col}\") for col in cols]\n expr_last = [pl.last(col).alias(f\"last_{col}\") for col in cols]\n #expr_first = [pl.first(col).alias(f\"first_{col}\") for col in cols]\n #expr_count = [pl.count(col).alias(f\"count_{col}\") for col in cols]\n return expr_max +expr_last#+expr_count\n \n def other_expr(df):\n cols = [col for col in df.columns if col[-1] in (\"T\", \"L\")]\n expr_max = [pl.max(col).alias(f\"max_{col}\") for col in cols]\n #expr_min = [pl.min(col).alias(f\"min_{col}\") for col in cols]\n expr_last = [pl.last(col).alias(f\"last_{col}\") for col in cols]\n #expr_first = [pl.first(col).alias(f\"first_{col}\") for col in cols]\n return expr_max +expr_last\n \n def count_expr(df):\n cols = [col for col in df.columns if \"num_group\" in col]\n expr_max = [pl.max(col).alias(f\"max_{col}\") for col in cols] \n #expr_min = [pl.min(col).alias(f\"min_{col}\") for col in cols]\n expr_last = [pl.last(col).alias(f\"last_{col}\") for col in cols]\n #expr_first = [pl.first(col).alias(f\"first_{col}\") for col in cols]\n return expr_max +expr_last\n \n def get_exprs(df):\n exprs = Aggregator.num_expr(df) + \\\n Aggregator.date_expr(df) + \\\n Aggregator.str_expr(df) + \\\n Aggregator.other_expr(df) + \\\n Aggregator.count_expr(df)\n\n return exprs","metadata":{"papermill":{"duration":0.027774,"end_time":"2024-04-02T05:49:57.017749","exception":false,"start_time":"2024-04-02T05:49:56.989975","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-04-08T11:34:09.969663Z","iopub.execute_input":"2024-04-08T11:34:09.970159Z","iopub.status.idle":"2024-04-08T11:34:09.999442Z","shell.execute_reply.started":"2024-04-08T11:34:09.970136Z","shell.execute_reply":"2024-04-08T11:34:09.998630Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def read_file(path, depth=None):\n df = pl.read_parquet(path)\n df = df.pipe(Pipeline.set_table_dtypes)\n if depth in [1,2]:\n df = df.group_by(\"case_id\").agg(Aggregator.get_exprs(df)) \n return df\n\ndef read_files(regex_path, depth=None):\n chunks = []\n \n for path in glob(str(regex_path)):\n df = pl.read_parquet(path)\n df = df.pipe(Pipeline.set_table_dtypes)\n if depth in [1, 2]:\n df = df.group_by(\"case_id\").agg(Aggregator.get_exprs(df))\n chunks.append(df)\n \n df = pl.concat(chunks, how=\"vertical_relaxed\")\n df = df.unique(subset=[\"case_id\"])\n return df","metadata":{"papermill":{"duration":0.020599,"end_time":"2024-04-02T05:49:57.049001","exception":false,"start_time":"2024-04-02T05:49:57.028402","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-04-08T11:34:10.001062Z","iopub.execute_input":"2024-04-08T11:34:10.001308Z","iopub.status.idle":"2024-04-08T11:34:10.015230Z","shell.execute_reply.started":"2024-04-08T11:34:10.001288Z","shell.execute_reply":"2024-04-08T11:34:10.014388Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def feature_eng(df_base, depth_0, depth_1, depth_2):\n df_base = (\n df_base\n .with_columns(\n month_decision = pl.col(\"date_decision\").dt.month(),\n weekday_decision = pl.col(\"date_decision\").dt.weekday(),\n )\n )\n for i, df in enumerate(depth_0 + depth_1 + depth_2):\n df_base = df_base.join(df, how=\"left\", on=\"case_id\", suffix=f\"_{i}\")\n df_base = df_base.pipe(Pipeline.handle_dates)\n return df_base","metadata":{"papermill":{"duration":0.019121,"end_time":"2024-04-02T05:49:57.079333","exception":false,"start_time":"2024-04-02T05:49:57.060212","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-04-08T11:34:10.016285Z","iopub.execute_input":"2024-04-08T11:34:10.016641Z","iopub.status.idle":"2024-04-08T11:34:10.027170Z","shell.execute_reply.started":"2024-04-08T11:34:10.016610Z","shell.execute_reply":"2024-04-08T11:34:10.026408Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def to_pandas(df_data, cat_cols=None):\n df_data = df_data.to_pandas()\n if cat_cols is None:\n cat_cols = list(df_data.select_dtypes(\"object\").columns)\n df_data[cat_cols] = df_data[cat_cols].astype(\"category\")\n return df_data, cat_cols","metadata":{"papermill":{"duration":0.018109,"end_time":"2024-04-02T05:49:57.108159","exception":false,"start_time":"2024-04-02T05:49:57.090050","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-04-08T11:34:10.028281Z","iopub.execute_input":"2024-04-08T11:34:10.028534Z","iopub.status.idle":"2024-04-08T11:34:10.038549Z","shell.execute_reply.started":"2024-04-08T11:34:10.028513Z","shell.execute_reply":"2024-04-08T11:34:10.037689Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def reduce_mem_usage(df):\n \"\"\" iterate through all the columns of a dataframe and modify the data type\n to reduce memory usage. \n \"\"\"\n start_mem = df.memory_usage().sum() / 1024**2\n print('Memory usage of dataframe is {:.2f} MB'.format(start_mem))\n \n for col in df.columns:\n col_type = df[col].dtype\n if str(col_type)==\"category\":\n continue\n \n if col_type != object:\n c_min = df[col].min()\n c_max = df[col].max()\n if str(col_type)[:3] == 'int':\n if c_min > np.iinfo(np.int8).min and c_max < np.iinfo(np.int8).max:\n df[col] = df[col].astype(np.int8)\n elif c_min > np.iinfo(np.int16).min and c_max < np.iinfo(np.int16).max:\n df[col] = df[col].astype(np.int16)\n elif c_min > np.iinfo(np.int32).min and c_max < np.iinfo(np.int32).max:\n df[col] = df[col].astype(np.int32)\n elif c_min > np.iinfo(np.int64).min and c_max < np.iinfo(np.int64).max:\n df[col] = df[col].astype(np.int64) \n else:\n if c_min > np.finfo(np.float16).min and c_max < np.finfo(np.float16).max:\n df[col] = df[col].astype(np.float16)\n elif c_min > np.finfo(np.float32).min and c_max < np.finfo(np.float32).max:\n df[col] = df[col].astype(np.float32)\n else:\n df[col] = df[col].astype(np.float64)\n else:\n continue\n end_mem = df.memory_usage().sum() / 1024**2\n print('Memory usage after optimization is: {:.2f} MB'.format(end_mem))\n print('Decreased by {:.1f}%'.format(100 * (start_mem - end_mem) / start_mem))\n \n return df","metadata":{"papermill":{"duration":0.025639,"end_time":"2024-04-02T05:49:57.144789","exception":false,"start_time":"2024-04-02T05:49:57.119150","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-04-08T11:34:10.039637Z","iopub.execute_input":"2024-04-08T11:34:10.039867Z","iopub.status.idle":"2024-04-08T11:34:10.052917Z","shell.execute_reply.started":"2024-04-08T11:34:10.039847Z","shell.execute_reply":"2024-04-08T11:34:10.051950Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def reduce_group(df, grps):\n use = []\n for g in grps:\n mx = 0; vx = g[0]\n for gg in g:\n n = df[gg].nunique()\n if n>mx:\n mx = n\n vx = gg\n #print(str(gg)+'-'+str(n),', ',end='')\n use.append(vx)\n #print()\n print('Use these',use)\n return use\n\ndef group_columns_by_correlation(matrix, threshold=0.7):\n # 莥įŽ—列䚋间įš„į›¸å…ŗ性\n correlation_matrix = matrix.corr()\n\n # 分įģ„列\n groups = []\n remaining_cols = list(matrix.columns)\n while remaining_cols:\n col = remaining_cols.pop(0)\n group = [col]\n correlated_cols = [col]\n for c in remaining_cols:\n if correlation_matrix.loc[col, c] >= threshold:\n group.append(c)\n correlated_cols.append(c)\n groups.append(group)\n remaining_cols = [c for c in remaining_cols if c not in correlated_cols]\n \n return groups","metadata":{"execution":{"iopub.status.busy":"2024-04-08T11:34:10.056332Z","iopub.execute_input":"2024-04-08T11:34:10.056572Z","iopub.status.idle":"2024-04-08T11:34:10.067565Z","shell.execute_reply.started":"2024-04-08T11:34:10.056552Z","shell.execute_reply":"2024-04-08T11:34:10.066826Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# ROOT = Path(\"home-credit-credit-risk-model-stability\")\nROOT = Path(\"/kaggle/input/home-credit-credit-risk-model-stability\")\n\nTRAIN_DIR = ROOT / \"parquet_files\" / \"train\"\nTEST_DIR = ROOT / \"parquet_files\" / \"test\"","metadata":{"papermill":{"duration":0.017298,"end_time":"2024-04-02T05:49:57.174569","exception":false,"start_time":"2024-04-02T05:49:57.157271","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-04-08T11:34:10.068526Z","iopub.execute_input":"2024-04-08T11:34:10.068802Z","iopub.status.idle":"2024-04-08T11:34:10.081850Z","shell.execute_reply.started":"2024-04-08T11:34:10.068781Z","shell.execute_reply":"2024-04-08T11:34:10.081008Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"data_store = {\n \"df_base\": read_file(TRAIN_DIR / \"train_base.parquet\"),\n \"depth_0\": [\n read_file(TRAIN_DIR / \"train_static_cb_0.parquet\"),\n read_files(TRAIN_DIR / \"train_static_0_*.parquet\"),\n ],\n \"depth_1\": [\n read_files(TRAIN_DIR / \"train_applprev_1_*.parquet\", 1),\n read_file(TRAIN_DIR / \"train_tax_registry_a_1.parquet\", 1),\n read_file(TRAIN_DIR / \"train_tax_registry_b_1.parquet\", 1),\n read_file(TRAIN_DIR / \"train_tax_registry_c_1.parquet\", 1),\n read_files(TRAIN_DIR / \"train_credit_bureau_a_1_*.parquet\", 1),\n read_file(TRAIN_DIR / \"train_credit_bureau_b_1.parquet\", 1),\n read_file(TRAIN_DIR / \"train_other_1.parquet\", 1),\n read_file(TRAIN_DIR / \"train_person_1.parquet\", 1),\n read_file(TRAIN_DIR / \"train_deposit_1.parquet\", 1),\n read_file(TRAIN_DIR / \"train_debitcard_1.parquet\", 1),\n ],\n \"depth_2\": [\n read_file(TRAIN_DIR / \"train_credit_bureau_b_2.parquet\", 2),\n read_files(TRAIN_DIR / \"train_credit_bureau_a_2_*.parquet\", 2),\n read_file(TRAIN_DIR / \"train_applprev_2.parquet\", 2),\n read_file(TRAIN_DIR / \"train_person_2.parquet\", 2)\n ]\n}","metadata":{"papermill":{"duration":140.566703,"end_time":"2024-04-02T05:52:17.751862","exception":false,"start_time":"2024-04-02T05:49:57.185159","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-04-08T11:34:10.082920Z","iopub.execute_input":"2024-04-08T11:34:10.083807Z","iopub.status.idle":"2024-04-08T11:36:27.118620Z","shell.execute_reply.started":"2024-04-08T11:34:10.083777Z","shell.execute_reply":"2024-04-08T11:36:27.117570Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df_train = feature_eng(**data_store)\nprint(\"train data shape:\\t\", df_train.shape)\ndel data_store\n","metadata":{"papermill":{"duration":23.064716,"end_time":"2024-04-02T05:52:40.828062","exception":false,"start_time":"2024-04-02T05:52:17.763346","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-04-08T11:36:27.119950Z","iopub.execute_input":"2024-04-08T11:36:27.120314Z","iopub.status.idle":"2024-04-08T11:36:43.249491Z","shell.execute_reply.started":"2024-04-08T11:36:27.120283Z","shell.execute_reply":"2024-04-08T11:36:43.248385Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df_train = Pipeline.filter_cols(df_train, 0.6)\nprint(\"after filteriing shape:\\t\", df_train.shape)\ngc.collect()","metadata":{"execution":{"iopub.status.busy":"2024-04-08T11:36:43.250632Z","iopub.execute_input":"2024-04-08T11:36:43.250960Z","iopub.status.idle":"2024-04-08T11:36:48.037352Z","shell.execute_reply.started":"2024-04-08T11:36:43.250933Z","shell.execute_reply":"2024-04-08T11:36:48.036457Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"encoding_cols = df_train.select(pl.selectors.by_dtype([pl.String, pl.Categorical])).columns\nfor col in encoding_cols:\n df_train = df_train.with_columns(pl.col(col).fill_null(\"Missing\"))","metadata":{"execution":{"iopub.status.busy":"2024-04-08T11:36:48.038630Z","iopub.execute_input":"2024-04-08T11:36:48.039055Z","iopub.status.idle":"2024-04-08T11:36:49.555177Z","shell.execute_reply.started":"2024-04-08T11:36:48.039015Z","shell.execute_reply":"2024-04-08T11:36:49.554163Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"data_store = {\n \"df_base\": read_file(TEST_DIR / \"test_base.parquet\"),\n \"depth_0\": [\n read_file(TEST_DIR / \"test_static_cb_0.parquet\"),\n read_files(TEST_DIR / \"test_static_0_*.parquet\"),\n ],\n \"depth_1\": [\n read_files(TEST_DIR / \"test_applprev_1_*.parquet\", 1),\n read_file(TEST_DIR / \"test_tax_registry_a_1.parquet\", 1),\n read_file(TEST_DIR / \"test_tax_registry_b_1.parquet\", 1),\n read_file(TEST_DIR / \"test_tax_registry_c_1.parquet\", 1),\n read_files(TEST_DIR / \"test_credit_bureau_a_1_*.parquet\", 1),\n read_file(TEST_DIR / \"test_credit_bureau_b_1.parquet\", 1),\n read_file(TEST_DIR / \"test_other_1.parquet\", 1),\n read_file(TEST_DIR / \"test_person_1.parquet\", 1),\n read_file(TEST_DIR / \"test_deposit_1.parquet\", 1),\n read_file(TEST_DIR / \"test_debitcard_1.parquet\", 1),\n ],\n \"depth_2\": [\n read_file(TEST_DIR / \"test_credit_bureau_b_2.parquet\", 2),\n read_files(TEST_DIR / \"test_credit_bureau_a_2_*.parquet\", 2),\n read_file(TEST_DIR / \"test_applprev_2.parquet\", 2),\n read_file(TEST_DIR / \"test_person_2.parquet\", 2)\n ]\n}","metadata":{"execution":{"iopub.status.busy":"2024-04-08T11:36:49.556412Z","iopub.execute_input":"2024-04-08T11:36:49.556746Z","iopub.status.idle":"2024-04-08T11:36:49.875339Z","shell.execute_reply.started":"2024-04-08T11:36:49.556720Z","shell.execute_reply":"2024-04-08T11:36:49.874397Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df_test = feature_eng(**data_store)\ntrain_cols = df_train.columns\ntrain_cols.remove('target')\nprint(\"test data shape:\\t\", df_test.shape)\ndel data_store\ndf_test = df_test.select(train_cols)\ndel train_cols\nprint(\"after filteriing shape:\\t\", df_test.shape)\n\ngc.collect()","metadata":{"execution":{"iopub.status.busy":"2024-04-08T11:36:49.876498Z","iopub.execute_input":"2024-04-08T11:36:49.877137Z","iopub.status.idle":"2024-04-08T11:36:50.070897Z","shell.execute_reply.started":"2024-04-08T11:36:49.877105Z","shell.execute_reply":"2024-04-08T11:36:50.069867Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"for col in encoding_cols:\n df_test = df_test.with_columns(pl.col(col).fill_null(\"Missing\"))","metadata":{"execution":{"iopub.status.busy":"2024-04-08T11:36:50.072165Z","iopub.execute_input":"2024-04-08T11:36:50.072861Z","iopub.status.idle":"2024-04-08T11:36:50.102368Z","shell.execute_reply.started":"2024-04-08T11:36:50.072836Z","shell.execute_reply":"2024-04-08T11:36:50.101512Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Count Encoding","metadata":{}},{"cell_type":"code","source":"# encoding_df = pl.concat([df_train.drop('target'), df_test])\n# cnt_mappings = {}\n# for col in encoding_cols:\n# cnt_mappings[col] = encoding_df.group_by(col).len()\n# del encoding_df\n\n# df_train_lazy = df_train.select(cnt_mappings.keys()).lazy()\n# df_test_lazy = df_test.select(cnt_mappings.keys()).lazy()\n\n# for col, mapping in cnt_mappings.items():\n# remapping = {category: count for category, count in mapping.rows()}\n# remapping[None] = -2\n# expr = pl.col(col).replace(\n# remapping,\n# default=-1,\n# )\n# df_train_lazy = df_train_lazy.with_columns(expr.alias(col + '_cnt'))\n# df_test_lazy = df_test_lazy.with_columns(expr.alias(col + '_cnt'))\n# del col, mapping, remapping\n# gc.collect()\n\n# transformed_train = df_train_lazy.collect()\n# transformed_test = df_test_lazy.collect()\n# del df_train_lazy, df_test_lazy\n\n# df_train = pl.concat([df_train, transformed_train.select(\"^*cnt$\")], how='horizontal')\n# df_test = pl.concat([df_test, transformed_test.select(\"^*cnt$\")], how='horizontal')\n# del transformed_train, transformed_test\n\n# gc.collect()\n","metadata":{"papermill":{"duration":22.596856,"end_time":"2024-04-02T05:53:03.437943","exception":false,"start_time":"2024-04-02T05:52:40.841087","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-04-08T11:36:50.103499Z","iopub.execute_input":"2024-04-08T11:36:50.104054Z","iopub.status.idle":"2024-04-08T11:37:13.935360Z","shell.execute_reply.started":"2024-04-08T11:36:50.104022Z","shell.execute_reply":"2024-04-08T11:37:13.934449Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df_train.shape","metadata":{"execution":{"iopub.status.busy":"2024-04-08T11:37:13.936373Z","iopub.execute_input":"2024-04-08T11:37:13.936666Z","iopub.status.idle":"2024-04-08T11:37:13.942368Z","shell.execute_reply.started":"2024-04-08T11:37:13.936642Z","shell.execute_reply":"2024-04-08T11:37:13.941461Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df_test.shape","metadata":{"execution":{"iopub.status.busy":"2024-04-08T11:37:13.943456Z","iopub.execute_input":"2024-04-08T11:37:13.943747Z","iopub.status.idle":"2024-04-08T11:37:13.954948Z","shell.execute_reply.started":"2024-04-08T11:37:13.943715Z","shell.execute_reply":"2024-04-08T11:37:13.954027Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# TargetEncoding","metadata":{}},{"cell_type":"code","source":"targetlist = np.ndarray((len(df_train)+len(df_test), ))\ntargetlist[:len(df_train)] = df_train['target'].to_numpy().flatten()\ntargetlist[len(df_train):] = None","metadata":{"execution":{"iopub.status.busy":"2024-04-08T11:37:13.956152Z","iopub.execute_input":"2024-04-08T11:37:13.956473Z","iopub.status.idle":"2024-04-08T11:37:14.052752Z","shell.execute_reply.started":"2024-04-08T11:37:13.956445Z","shell.execute_reply":"2024-04-08T11:37:14.052009Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"targetlist","metadata":{"execution":{"iopub.status.busy":"2024-04-08T11:37:14.053793Z","iopub.execute_input":"2024-04-08T11:37:14.054121Z","iopub.status.idle":"2024-04-08T11:37:14.061644Z","shell.execute_reply.started":"2024-04-08T11:37:14.054093Z","shell.execute_reply":"2024-04-08T11:37:14.060626Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"encoding_df = pl.concat([df_train.drop('target'), df_test])\nencoding_df = encoding_df.with_columns(pl.Series(targetlist).alias('target'))\n\ntarget_mappings = {}\nfor col in encoding_cols:\n target_mappings[col] = encoding_df.group_by(col).agg(pl.col('target').mean())\n\ndf_train_lazy = df_train.select(target_mappings.keys()).lazy()\ndf_test_lazy = df_test.select(target_mappings.keys()).lazy()\n\nfor col, mapping in target_mappings.items():\n remapping = {category: count for category, count in mapping.rows()}\n remapping[None] = -2\n expr = pl.col(col).replace(\n remapping,\n default=-1,\n )\n df_train_lazy = df_train_lazy.with_columns(expr.alias(col + '_targetEncoding'))\n df_test_lazy = df_test_lazy.with_columns(expr.alias(col + '_targetEncoding'))\n del col, mapping, remapping\n gc.collect()\n\ntransformed_train = df_train_lazy.collect()\ntransformed_test = df_test_lazy.collect()\ndel df_train_lazy, df_test_lazy\n\ndf_train = pl.concat([df_train, transformed_train.select(\"^*_targetEncoding$\")], how='horizontal')\ndf_test = pl.concat([df_test, transformed_test.select(\"^*_targetEncoding$\")], how='horizontal')\ndel transformed_train, transformed_test, encoding_df\n\ngc.collect()\n","metadata":{"execution":{"iopub.status.busy":"2024-04-08T11:37:14.062464Z","iopub.execute_input":"2024-04-08T11:37:14.062759Z","iopub.status.idle":"2024-04-08T11:37:37.787254Z","shell.execute_reply.started":"2024-04-08T11:37:14.062737Z","shell.execute_reply":"2024-04-08T11:37:37.786415Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"\ndf_train, cat_cols = to_pandas(df_train)\ndf_train = reduce_mem_usage(df_train)\nprint(\"train data shape:\\t\", df_train.shape)\n\ndf_test, cat_cols = to_pandas(df_test, cat_cols)\ndf_test = reduce_mem_usage(df_test)\nprint(\"test data shape:\\t\", df_test.shape)\ngc.collect()\n","metadata":{"execution":{"iopub.status.busy":"2024-04-08T11:37:37.788237Z","iopub.execute_input":"2024-04-08T11:37:37.788483Z","iopub.status.idle":"2024-04-08T11:38:21.015019Z","shell.execute_reply.started":"2024-04-08T11:37:37.788462Z","shell.execute_reply":"2024-04-08T11:38:21.014021Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"nums=df_train.select_dtypes(exclude='category').columns\n\n#df_train=df_train[nums]\nnans_df = df_train[nums].isna()\nnans_groups={}\nfor col in nums:\n cur_group = nans_df[col].sum()\n try:\n nans_groups[cur_group].append(col)\n except:\n nans_groups[cur_group]=[col]\ndel nans_df; x=gc.collect()\n\nuses=[]\nfor k,v in nans_groups.items():\n if len(v)>1:\n Vs = nans_groups[k]\n #cross_features=list(combinations(Vs, 2))\n #make_corr(Vs)\n grps= group_columns_by_correlation(df_train[Vs], threshold=0.8)\n use=reduce_group(df_train, grps)\n uses=uses+use\n #make_corr(use)\n else:\n uses=uses+v\n print('####### NAN count =',k)\nprint(uses)\nprint(len(uses))\nuses=uses+list(df_train.select_dtypes(include='category').columns)\nprint(len(uses))\ndf_train=df_train[uses]\ndf_train.shape","metadata":{"papermill":{"duration":172.403171,"end_time":"2024-04-02T05:55:55.852785","exception":false,"start_time":"2024-04-02T05:53:03.449614","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-04-08T11:38:21.016178Z","iopub.execute_input":"2024-04-08T11:38:21.016485Z","iopub.status.idle":"2024-04-08T11:40:16.233923Z","shell.execute_reply.started":"2024-04-08T11:38:21.016461Z","shell.execute_reply":"2024-04-08T11:40:16.232970Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_cols = df_train.columns.to_list()\ntrain_cols.remove('target')\ndf_test = df_test[train_cols]","metadata":{"execution":{"iopub.status.busy":"2024-04-08T11:40:16.239234Z","iopub.execute_input":"2024-04-08T11:40:16.239532Z","iopub.status.idle":"2024-04-08T11:40:16.264074Z","shell.execute_reply.started":"2024-04-08T11:40:16.239507Z","shell.execute_reply":"2024-04-08T11:40:16.263198Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"sample = pd.read_csv(\"/kaggle/input/home-credit-credit-risk-model-stability/sample_submission.csv\")\n# sample = pd.read_csv(\"home-credit-credit-risk-model-stability/sample_submission.csv\")\ndevice='gpu'\n#n_samples=200000\nDRY_RUN = True if sample.shape[0] == 10 else False \nif DRY_RUN:\n device='cpu'\n df_train = df_train.iloc[:100000, :]\nprint(device)","metadata":{"papermill":{"duration":0.029689,"end_time":"2024-04-02T05:55:55.896892","exception":false,"start_time":"2024-04-02T05:55:55.867203","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-04-08T11:40:16.265303Z","iopub.execute_input":"2024-04-08T11:40:16.265726Z","iopub.status.idle":"2024-04-08T11:40:16.307217Z","shell.execute_reply.started":"2024-04-08T11:40:16.265694Z","shell.execute_reply":"2024-04-08T11:40:16.306390Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# ## no work (too slow..) \n# ## require dimensionality reduction first & features with high feature impor. \n# imputer = KNNImputer()\n# df_train = imputer.fit_transform(df_train)","metadata":{"papermill":{"duration":0.021361,"end_time":"2024-04-02T05:55:57.446276","exception":false,"start_time":"2024-04-02T05:55:57.424915","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-04-08T11:40:16.308398Z","iopub.execute_input":"2024-04-08T11:40:16.308751Z","iopub.status.idle":"2024-04-08T11:40:16.312995Z","shell.execute_reply.started":"2024-04-08T11:40:16.308720Z","shell.execute_reply":"2024-04-08T11:40:16.312001Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"convert_cols = ['month_decision', 'weekday_decision']\n\ndf_train[convert_cols] = df_train[convert_cols].astype('category')\ndf_test[convert_cols] = df_test[convert_cols].astype('category')","metadata":{"execution":{"iopub.status.busy":"2024-04-08T11:40:16.314029Z","iopub.execute_input":"2024-04-08T11:40:16.314297Z","iopub.status.idle":"2024-04-08T11:40:16.334830Z","shell.execute_reply.started":"2024-04-08T11:40:16.314275Z","shell.execute_reply":"2024-04-08T11:40:16.333903Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(df_train.shape)\nprint(df_test.shape)","metadata":{"execution":{"iopub.status.busy":"2024-04-08T11:40:16.335871Z","iopub.execute_input":"2024-04-08T11:40:16.336129Z","iopub.status.idle":"2024-04-08T11:40:16.340620Z","shell.execute_reply.started":"2024-04-08T11:40:16.336107Z","shell.execute_reply":"2024-04-08T11:40:16.339745Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"y = df_train[\"target\"]\nweeks = df_train[\"WEEK_NUM\"]\ndf_train= df_train.drop(columns=[\"target\", \"case_id\", \"WEEK_NUM\"])\n\ncv = StratifiedGroupKFold(n_splits=5, shuffle=False)","metadata":{"papermill":{"duration":1.231344,"end_time":"2024-04-02T05:55:58.692179","exception":false,"start_time":"2024-04-02T05:55:57.460835","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-04-08T11:40:16.341727Z","iopub.execute_input":"2024-04-08T11:40:16.341974Z","iopub.status.idle":"2024-04-08T11:40:16.744967Z","shell.execute_reply.started":"2024-04-08T11:40:16.341952Z","shell.execute_reply":"2024-04-08T11:40:16.743924Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df_train","metadata":{"execution":{"iopub.status.busy":"2024-04-08T11:40:16.746083Z","iopub.execute_input":"2024-04-08T11:40:16.746346Z","iopub.status.idle":"2024-04-08T11:40:16.802237Z","shell.execute_reply.started":"2024-04-08T11:40:16.746325Z","shell.execute_reply":"2024-04-08T11:40:16.801370Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"\nparams = {\n \"boosting_type\": \"gbdt\",\n \"objective\": \"binary\",\n \"metric\": \"auc\",\n \"max_depth\": 10, \n \"learning_rate\": 0.05,\n \"n_estimators\": 2000, \n \"colsample_bytree\": 0.8,\n \"colsample_bynode\": 0.8,\n \"verbose\": -1,\n \"random_state\": 42,\n \"reg_alpha\": 0.1,\n \"reg_lambda\": 10,\n \"extra_trees\":True,\n 'num_leaves':64,\n # \"sample_weight\":'balanced',\n 'categorical_feature': 'auto',\n \"device\": device, \n \"verbose\": -1,\n}\n\nfitted_models = []\ncv_scores = []\n\n\nfor idx_train, idx_valid in cv.split(df_train, y, groups=weeks):# Because it takes a long time to divide the data set, \n X_train, y_train = df_train.iloc[idx_train], y.iloc[idx_train]# each time the data set is divided, two models are trained to each other twice, which saves time.\n X_valid, y_valid = df_train.iloc[idx_valid], y.iloc[idx_valid]\n model = lgb.LGBMClassifier(**params)\n model.fit(\n X_train, y_train,\n eval_set = [(X_valid, y_valid)],\n callbacks = [lgb.early_stopping(100)] )\n fitted_models.append(model)\n y_pred_valid = model.predict_proba(X_valid)[:,1]\n auc_score = roc_auc_score(y_valid, y_pred_valid)\n cv_scores.append(auc_score)\n \nprint(\"CV AUC scores: \", cv_scores)\nprint(\"Maximum CV AUC score: \", max(cv_scores))","metadata":{"papermill":{"duration":1783.036267,"end_time":"2024-04-02T06:25:41.743530","exception":false,"start_time":"2024-04-02T05:55:58.707263","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-04-08T11:40:16.803409Z","iopub.execute_input":"2024-04-08T11:40:16.803782Z","iopub.status.idle":"2024-04-08T11:51:37.133456Z","shell.execute_reply.started":"2024-04-08T11:40:16.803749Z","shell.execute_reply":"2024-04-08T11:51:37.132433Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"class VotingModel(BaseEstimator, RegressorMixin):\n def __init__(self, estimators):\n super().__init__()\n self.estimators = estimators\n \n def fit(self, X, y=None):\n return self\n \n def predict(self, X):\n y_preds = [estimator.predict(X) for estimator in self.estimators]\n return np.mean(y_preds, axis=0)\n \n def predict_proba(self, X):\n y_preds = [estimator.predict_proba(X) for estimator in self.estimators]\n return np.mean(y_preds, axis=0)\n\nmodel = VotingModel(fitted_models)","metadata":{"papermill":{"duration":0.03076,"end_time":"2024-04-02T06:25:41.796920","exception":false,"start_time":"2024-04-02T06:25:41.766160","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-04-08T11:51:37.134883Z","iopub.execute_input":"2024-04-08T11:51:37.135604Z","iopub.status.idle":"2024-04-08T11:51:37.142874Z","shell.execute_reply.started":"2024-04-08T11:51:37.135553Z","shell.execute_reply":"2024-04-08T11:51:37.141983Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# lgb.plot_importance(fitted_models[2], importance_type=\"split\", figsize=(10,50))\n# plt.show()","metadata":{"papermill":{"duration":5.882693,"end_time":"2024-04-02T06:25:47.700370","exception":false,"start_time":"2024-04-02T06:25:41.817677","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-04-08T11:51:37.144219Z","iopub.execute_input":"2024-04-08T11:51:37.145121Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"features = X_train.columns\nimportances = fitted_models[2].feature_importances_\nfeature_importance = pd.DataFrame({'importance':importances,'features':features}).sort_values('importance', ascending=False).reset_index(drop=True)\nfeature_importance\n\ndrop_list = []\nfor i, f in feature_importance.iterrows():\n if f['importance']<80:\n drop_list.append(f['features'])\nprint(f\"Number of features which are not important: {len(drop_list)} \")\n\nprint(drop_list)","metadata":{"papermill":{"duration":0.093075,"end_time":"2024-04-02T06:25:47.845894","exception":false,"start_time":"2024-04-02T06:25:47.752819","status":"completed"},"tags":[],"execution":{"iopub.execute_input":"2024-04-08T11:51:42.738333Z","iopub.status.idle":"2024-04-08T11:51:42.784351Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### Submission","metadata":{"papermill":{"duration":0.04982,"end_time":"2024-04-02T06:25:47.947360","exception":false,"start_time":"2024-04-02T06:25:47.897540","status":"completed"},"tags":[]}},{"cell_type":"code","source":"week_num = df_test['WEEK_NUM'].to_list()\n\ndf_test.drop([\"WEEK_NUM\"], axis=1, inplace=True)\ndf_test = df_test.set_index(\"case_id\")\n\ny_pred = pd.Series(model.predict_proba(df_test)[:, 1], index=df_test.index)\ndf_subm = pd.read_csv(ROOT / \"sample_submission.csv\")\ndf_subm = df_subm.set_index(\"case_id\")\ndf_subm['WEEK_NUM'] = week_num\ndf_subm[\"score\"] = y_pred","metadata":{"papermill":{"duration":0.080925,"end_time":"2024-04-02T06:25:48.079451","exception":false,"start_time":"2024-04-02T06:25:47.998526","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-04-08T11:51:42.785699Z","iopub.execute_input":"2024-04-08T11:51:42.786019Z","iopub.status.idle":"2024-04-08T11:51:43.238929Z","shell.execute_reply.started":"2024-04-08T11:51:42.785993Z","shell.execute_reply":"2024-04-08T11:51:43.237750Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"SHIFT = 0.025\n\ncondition = df_subm[\"WEEK_NUM\"] < (df_subm[\"WEEK_NUM\"].max() - df_subm[\"WEEK_NUM\"].min())/2 + df_subm[\"WEEK_NUM\"].min()\ndf_subm.loc[condition, 'score'] = (df_subm.loc[condition, 'score'] - SHIFT).clip(0)\ndel df_subm['WEEK_NUM']\ndf_subm.to_csv(\"submission.csv\")\ndf_subm","metadata":{"papermill":{"duration":0.900897,"end_time":"2024-04-02T06:25:49.032489","exception":false,"start_time":"2024-04-02T06:25:48.131592","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-04-08T11:51:43.240253Z","iopub.execute_input":"2024-04-08T11:51:43.240649Z","iopub.status.idle":"2024-04-08T11:51:43.258238Z","shell.execute_reply.started":"2024-04-08T11:51:43.240617Z","shell.execute_reply":"2024-04-08T11:51:43.257441Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{"papermill":{"duration":0.051392,"end_time":"2024-04-02T06:25:49.672425","exception":false,"start_time":"2024-04-02T06:25:49.621033","status":"completed"},"tags":[]},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]}]}