{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"name":"python","version":"3.10.12","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"nvidiaTeslaT4","dataSources":[{"sourceId":50160,"databundleVersionId":7602123,"sourceType":"competition"}],"dockerImageVersionId":30635,"isInternetEnabled":false,"language":"python","sourceType":"notebook","isGpuEnabled":true}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"import polars as pl\nimport numpy as np\nimport pandas as pd\nimport lightgbm as lgb\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.metrics import roc_auc_score \n\nimport polars.selectors as cs\nimport matplotlib.pyplot as plt\nimport seaborn as sns\nimport missingno as mn\n\nfrom sklearn.model_selection import StratifiedGroupKFold\nfrom sklearn.base import BaseEstimator, RegressorMixin\n\nimport warnings\nwarnings.filterwarnings('ignore')\n\ndataPath = \"/kaggle/input/home-credit-credit-risk-model-stability/\"","metadata":{"execution":{"iopub.status.busy":"2024-02-23T03:02:57.512407Z","iopub.execute_input":"2024-02-23T03:02:57.513194Z","iopub.status.idle":"2024-02-23T03:02:57.519408Z","shell.execute_reply.started":"2024-02-23T03:02:57.513158Z","shell.execute_reply":"2024-02-23T03:02:57.518245Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def set_table_dtypes(df: pl.DataFrame) -> pl.DataFrame:\n    # implement here all desired dtypes for tables\n    # the following is just an example\n    for col in df.columns:\n        # last letter of column name will help you determine the type\n        if col[-1] in (\"P\", \"A\"):\n            df = df.with_columns(pl.col(col).cast(pl.Float64).alias(col))\n    return df\n\ndef convert_strings(df: pd.DataFrame) -> pd.DataFrame:\n    for col in df.columns:  \n        if df[col].dtype.name in ['object', 'string']:\n            df[col] = df[col].astype(\"string\").astype('category')\n            current_categories = df[col].cat.categories\n            new_categories = current_categories.to_list() + [\"Unknown\"]\n            new_dtype = pd.CategoricalDtype(categories=new_categories, ordered=True)\n            df[col] = df[col].astype(new_dtype)\n    return df","metadata":{"execution":{"iopub.status.busy":"2024-02-23T03:02:57.677811Z","iopub.execute_input":"2024-02-23T03:02:57.678147Z","iopub.status.idle":"2024-02-23T03:02:57.685870Z","shell.execute_reply.started":"2024-02-23T03:02:57.678119Z","shell.execute_reply":"2024-02-23T03:02:57.684812Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Load feature_def","metadata":{}},{"cell_type":"code","source":"def get_feature_definitions(columns):\n    return pl.DataFrame({'Variable': columns}).join(\n        feature_def,\n        on = 'Variable',\n        how = 'left',\n    )\n\nfeature_def = pl.read_csv(dataPath + \"feature_definitions.csv\")","metadata":{"execution":{"iopub.status.busy":"2024-02-23T03:02:57.976497Z","iopub.execute_input":"2024-02-23T03:02:57.977060Z","iopub.status.idle":"2024-02-23T03:02:57.985133Z","shell.execute_reply.started":"2024-02-23T03:02:57.977024Z","shell.execute_reply":"2024-02-23T03:02:57.984226Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Load Basetable","metadata":{}},{"cell_type":"code","source":"# train\ntrain_basetable = pl.read_csv(dataPath + \"csv_files/train/train_base.csv\")\n\n# test\ntest_basetable = pl.read_csv(dataPath + \"csv_files/test/test_base.csv\")","metadata":{"execution":{"iopub.status.busy":"2024-02-23T03:02:58.282760Z","iopub.execute_input":"2024-02-23T03:02:58.283467Z","iopub.status.idle":"2024-02-23T03:02:58.470988Z","shell.execute_reply.started":"2024-02-23T03:02:58.283426Z","shell.execute_reply":"2024-02-23T03:02:58.470028Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Check Basetable","metadata":{}},{"cell_type":"code","source":"# print(train_basetable.columns)\n# print('\\n')\n# print(train_basetable.shape)\n# print('\\n')\n# display(train_basetable.head())","metadata":{"execution":{"iopub.status.busy":"2024-02-23T03:02:58.584550Z","iopub.execute_input":"2024-02-23T03:02:58.585339Z","iopub.status.idle":"2024-02-23T03:02:58.589255Z","shell.execute_reply.started":"2024-02-23T03:02:58.585303Z","shell.execute_reply":"2024-02-23T03:02:58.588251Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Load depth=0","metadata":{}},{"cell_type":"code","source":"# train\ntrain_static = pl.concat(\n    [\n        pl.read_csv(dataPath + \"csv_files/train/train_static_0_0.csv\").pipe(set_table_dtypes),\n        pl.read_csv(dataPath + \"csv_files/train/train_static_0_1.csv\").pipe(set_table_dtypes),\n    ],\n    how=\"vertical_relaxed\",\n)\ntrain_static_cb = pl.read_csv(dataPath + \"csv_files/train/train_static_cb_0.csv\").pipe(set_table_dtypes)\n\n# test\ntest_static = pl.concat(\n    [\n        pl.read_csv(dataPath + \"csv_files/test/test_static_0_0.csv\").pipe(set_table_dtypes),\n        pl.read_csv(dataPath + \"csv_files/test/test_static_0_1.csv\").pipe(set_table_dtypes),\n        pl.read_csv(dataPath + \"csv_files/test/test_static_0_2.csv\").pipe(set_table_dtypes),\n    ],\n    how=\"vertical_relaxed\",\n)\ntest_static_cb = pl.read_csv(dataPath + \"csv_files/test/test_static_cb_0.csv\").pipe(set_table_dtypes)","metadata":{"execution":{"iopub.status.busy":"2024-02-23T03:02:58.889397Z","iopub.execute_input":"2024-02-23T03:02:58.889884Z","iopub.status.idle":"2024-02-23T03:03:09.496747Z","shell.execute_reply.started":"2024-02-23T03:02:58.889825Z","shell.execute_reply":"2024-02-23T03:03:09.495933Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Check depth=0","metadata":{}},{"cell_type":"code","source":"# print(train_static.columns)\n# print('\\n')\n# print(train_static.shape)\n# print('\\n')\n# display(train_static.head())","metadata":{"execution":{"iopub.status.busy":"2024-02-23T03:03:09.498472Z","iopub.execute_input":"2024-02-23T03:03:09.498726Z","iopub.status.idle":"2024-02-23T03:03:09.504954Z","shell.execute_reply.started":"2024-02-23T03:03:09.498704Z","shell.execute_reply":"2024-02-23T03:03:09.504063Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# print(train_static_cb.columns)\n# print('\\n')\n# print(train_static_cb.shape)\n# print('\\n')\n# display(train_static_cb.head())","metadata":{"execution":{"iopub.status.busy":"2024-02-23T03:03:09.506055Z","iopub.execute_input":"2024-02-23T03:03:09.506317Z","iopub.status.idle":"2024-02-23T03:03:09.536630Z","shell.execute_reply.started":"2024-02-23T03:03:09.506293Z","shell.execute_reply":"2024-02-23T03:03:09.535865Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# depth=1에 해당하는 테이블\n\nInternal file\n> `applprev_1` `debitcard_1` `deposit_1` `other_1` `person_1`\n\nExternal file\n> `credit_bureau_b_1` ","metadata":{}},{"cell_type":"code","source":"## applprev_1\n# train\ntrain_applprev_1 = pl.concat(\n    [\n        pl.read_csv(dataPath + \"csv_files/train/train_applprev_1_0.csv\").pipe(set_table_dtypes),\n        pl.read_csv(dataPath + 'csv_files/train/train_applprev_1_1.csv').pipe(set_table_dtypes),\n    ],\n    how='vertical_relaxed',\n)\n\n# test\ntest_applprev_1 = pl.concat(\n    [\n        pl.read_csv(dataPath + \"csv_files/test/test_applprev_1_0.csv\").pipe(set_table_dtypes),\n        pl.read_csv(dataPath + 'csv_files/test/test_applprev_1_1.csv').pipe(set_table_dtypes),\n        pl.read_csv(dataPath + 'csv_files/test/test_applprev_1_2.csv').pipe(set_table_dtypes),\n    ],\n    how='vertical_relaxed',\n)","metadata":{"execution":{"iopub.status.busy":"2024-02-23T03:03:09.538587Z","iopub.execute_input":"2024-02-23T03:03:09.538870Z","iopub.status.idle":"2024-02-23T03:03:17.668797Z","shell.execute_reply.started":"2024-02-23T03:03:09.538846Z","shell.execute_reply":"2024-02-23T03:03:17.667992Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# train\ntrain_debitcard_1 = pl.read_csv(dataPath + \"csv_files/train/train_debitcard_1.csv\").pipe(set_table_dtypes) \ntrain_deposit_1 = pl.read_csv(dataPath + \"csv_files/train/train_deposit_1.csv\").pipe(set_table_dtypes)\ntrain_other_1 = pl.read_csv(dataPath + 'csv_files/train/train_other_1.csv').pipe(set_table_dtypes)\ntrain_person_1 = pl.read_csv(dataPath + \"csv_files/train/train_person_1.csv\").pipe(set_table_dtypes)\ntrain_credit_bureau_b_1 = pl.read_csv(dataPath + 'csv_files/train/train_credit_bureau_b_1.csv').pipe(set_table_dtypes)\n\n# test\ntest_debitcard_1 = pl.read_csv(dataPath + \"csv_files/test/test_debitcard_1.csv\").pipe(set_table_dtypes)\ntest_deposit_1 = pl.read_csv(dataPath + \"csv_files/test/test_deposit_1.csv\").pipe(set_table_dtypes)\ntest_other_1 = pl.read_csv(dataPath + 'csv_files/test/test_other_1.csv').pipe(set_table_dtypes)\ntest_person_1 = pl.read_csv(dataPath + \"csv_files/test/test_person_1.csv\").pipe(set_table_dtypes)\ntest_credit_bureau_b_1 = pl.read_csv(dataPath + 'csv_files/test/test_credit_bureau_b_1.csv').pipe(set_table_dtypes)","metadata":{"execution":{"iopub.status.busy":"2024-02-23T03:03:17.669950Z","iopub.execute_input":"2024-02-23T03:03:17.670243Z","iopub.status.idle":"2024-02-23T03:03:20.597188Z","shell.execute_reply.started":"2024-02-23T03:03:17.670217Z","shell.execute_reply":"2024-02-23T03:03:20.596232Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(\n    train_applprev_1.shape,\n    train_debitcard_1.shape, \n    train_deposit_1.shape, \n    train_other_1.shape, \n    train_person_1.shape,\n    train_credit_bureau_b_1.shape,\n)","metadata":{"execution":{"iopub.status.busy":"2024-02-23T03:03:20.598257Z","iopub.execute_input":"2024-02-23T03:03:20.599147Z","iopub.status.idle":"2024-02-23T03:03:20.605785Z","shell.execute_reply.started":"2024-02-23T03:03:20.599105Z","shell.execute_reply":"2024-02-23T03:03:20.604939Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Feature Engineering","metadata":{}},{"cell_type":"code","source":"## applprev_1\napp_cols = [\n    'case_id',\n    'maxdpdtolerance_577P',\n    'employedfrom_700D',\n    'pmtnum_8L',\n    'creationdate_885D',\n    'credamount_590A',\n    'mainoccupationinc_437A',\n    'firstnonzeroinstldate_307D',\n    'dtlastpmtallstes_3545839D',\n    'annuity_853A',\n    'familystate_726L',\n    'approvaldate_319D',\n    'district_544M',\n    'dateactivated_425D',\n    'education_1138M',\n]\n\n# train\ntrain_applprev_1 = train_applprev_1[app_cols]\ntrain_applprev_1 = train_applprev_1.unique(subset='case_id', keep='first')\n\n# test\ntest_applprev_1 = test_applprev_1[app_cols]\ntest_applprev_1 = test_applprev_1.unique(subset='case_id', keep='first')","metadata":{"execution":{"iopub.status.busy":"2024-02-23T03:03:20.606874Z","iopub.execute_input":"2024-02-23T03:03:20.608097Z","iopub.status.idle":"2024-02-23T03:03:21.579687Z","shell.execute_reply.started":"2024-02-23T03:03:20.608061Z","shell.execute_reply":"2024-02-23T03:03:21.578834Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"## case_id를 기준으로 그룹화한 후, aggregation functions\n\n# debitcard_1\ntrain_debitcard_1_feats = train_debitcard_1.group_by(\"case_id\").agg(\n    # 체크카드의 평균 잔액의 최대값\n    pl.col('last180dayaveragebalance_704A').max().alias('max_last180dayaveragebalance_704A'),\n    # 180일간의 체크카드 거래액의 최대값\n    pl.col('last180dayturnover_1134A').max().alias('max_last180dayturnover_1134A'),\n    # 30일간의 체크카드 거래액의 최대값\n    pl.col(\"last30dayturnover_651A\").max().alias(\"max_last30dayturnover_651A\"),\n).sort(by='case_id')\n\n# deposit_1\ntrain_deposit_1_feats = train_deposit_1.group_by('case_id').agg(\n    # 예/적금의 합계\n    pl.sum('amount_416A').alias('sum_amount_416A'),\n).sort(by='case_id')\n\n# other_1\ntrain_other_1_feats = train_other_1.group_by(\"case_id\").agg(\n    # 입금액의 합계\n    pl.col(\"amtdepositincoming_4809444A\").sum().alias(\"sum_amtdepositincoming_4809444A\"),\n    # 출금액의 합계\n    pl.col(\"amtdepositoutgoing_4809442A\").sum().alias(\"sum_amtdepositoutgoing_4809442A\"),\n).sort(by='case_id')\n\n# person_1\ntrain_person_1_feats_1 = train_person_1.select(['case_id', 'incometype_1044T','num_group1']).filter(\n    pl.col('num_group1') == 0\n).drop('num_group1')\n\ntrain_person_1_feats_2 = train_person_1.select(['case_id', 'language1_981M','num_group1']).filter(\n    pl.col('num_group1') == 0\n).drop('num_group1')\n\n# credit_bureau_b_1\ntrain_credit_bureau_b_1_feats = train_credit_bureau_b_1.group_by('case_id').agg(\n    pl.col('installmentamount_644A').mean().alias('mean_installmentamount_644A'),\n    pl.col('maxdebtpduevalodued_3940955A').mean().alias('mean_maxdebtpduevalodued_3940955A'),\n    pl.col('overdueamountmax_950A').mean().alias('mean_overdueamountmax_950A'),\n    pl.col('residualamount_1093A').mean().alias('mean_residualamount_1093A'),\n).sort(by='case_id')\n\n\n## ⭐키워드 6가지 중 어떤걸 선택할지?\nselected_static_cols = []\nfor col in train_static.columns:\n    if col[-1] in (\"A\", \"D\", \"M\"):\n        selected_static_cols.append(col)\nprint(selected_static_cols)\n\nselected_static_cb_cols = []\nfor col in train_static_cb.columns:\n    if col[-1] in (\"A\", \"D\", \"M\"):\n        selected_static_cb_cols.append(col)\nprint(selected_static_cb_cols)\n\n\n## Join\ndata = train_basetable.join(\n    train_static.select([\"case_id\"]+selected_static_cols), how=\"left\", on=\"case_id\"\n).join(\n    train_static_cb.select([\"case_id\"]+selected_static_cb_cols), how=\"left\", on=\"case_id\"\n).join(\n    train_applprev_1, how=\"left\", on=\"case_id\"\n).join(\n    train_debitcard_1_feats, how=\"left\", on=\"case_id\"\n).join(\n    train_deposit_1_feats, how=\"left\", on=\"case_id\"\n).join(\n    train_other_1_feats, how=\"left\", on=\"case_id\"\n).join(\n    train_person_1_feats_1, how=\"left\", on=\"case_id\"\n).join(\n    train_person_1_feats_2, how=\"left\", on=\"case_id\"\n).join(\n    train_credit_bureau_b_1_feats, how=\"left\", on=\"case_id\"\n)\n\n\n## num_group1이 0인 경우 대출을 신청한 사람\n# train_debitcard_1_feats_2 = train_debitcard_1.select([\"case_id\", \"num_group1\"]).filter(\n#     pl.col(\"num_group1\") == 0 # 필터링 후\n# ).drop(\"num_group1\") # 컬럼 삭제","metadata":{"execution":{"iopub.status.busy":"2024-02-23T03:03:21.580946Z","iopub.execute_input":"2024-02-23T03:03:21.581198Z","iopub.status.idle":"2024-02-23T03:03:24.007348Z","shell.execute_reply.started":"2024-02-23T03:03:21.581176Z","shell.execute_reply":"2024-02-23T03:03:24.006573Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"test_debitcard_1_feats = test_debitcard_1.group_by(\"case_id\").agg(\n    pl.col('last180dayaveragebalance_704A').max().alias('max_last180dayaveragebalance_704A'),\n    pl.col('last180dayturnover_1134A').max().alias('max_last180dayturnover_1134A'),\n    pl.col(\"last30dayturnover_651A\").max().alias(\"max_last30dayturnover_651A\"),\n).sort(by='case_id')\n\ntest_deposit_1_feats = test_deposit_1.group_by('case_id').agg(\n    pl.sum('amount_416A').alias('sum_amount_416A')\n).sort(by='case_id')\n\ntest_other_1_feats = test_other_1.group_by(\"case_id\").agg(\n    pl.col(\"amtdepositincoming_4809444A\").sum().alias(\"sum_amtdepositincoming_4809444A\"),\n    pl.col(\"amtdepositoutgoing_4809442A\").sum().alias(\"sum_amtdepositoutgoing_4809442A\"),\n).sort(by='case_id')\n\ntest_person_1_feats_1 = test_person_1.select(['case_id', 'incometype_1044T','num_group1']).filter(\n    pl.col('num_group1') == 0\n).drop('num_group1')\n\ntest_person_1_feats_2 = test_person_1.select(['case_id', 'language1_981M','num_group1']).filter(\n    pl.col('num_group1') == 0\n).drop('num_group1')\n\ntest_credit_bureau_b_1_feats = test_credit_bureau_b_1.group_by('case_id').agg(\n    pl.col('installmentamount_644A').mean().alias('mean_installmentamount_644A'),\n    pl.col('maxdebtpduevalodued_3940955A').mean().alias('mean_maxdebtpduevalodued_3940955A'),\n    pl.col('overdueamountmax_950A').mean().alias('mean_overdueamountmax_950A'),\n    pl.col('residualamount_1093A').mean().alias('mean_residualamount_1093A'),\n).sort(by='case_id')\n\n\ndata_submission = test_basetable.join(\n    test_static.select([\"case_id\"]+selected_static_cols), how=\"left\", on=\"case_id\"\n).join(\n    test_static_cb.select([\"case_id\"]+selected_static_cb_cols), how=\"left\", on=\"case_id\"\n).join(\n    test_applprev_1, how=\"left\", on=\"case_id\"\n).join(\n    test_debitcard_1_feats, how=\"left\", on=\"case_id\"\n).join(\n    test_deposit_1_feats, how=\"left\", on=\"case_id\"\n).join(\n    test_other_1_feats, how=\"left\", on=\"case_id\"\n).join(\n    test_person_1_feats_1, how=\"left\", on=\"case_id\"\n).join(\n    test_person_1_feats_2, how=\"left\", on=\"case_id\"\n).join(\n    test_credit_bureau_b_1_feats, how=\"left\", on=\"case_id\"\n)","metadata":{"execution":{"iopub.status.busy":"2024-02-23T03:03:24.008673Z","iopub.execute_input":"2024-02-23T03:03:24.009283Z","iopub.status.idle":"2024-02-23T03:03:24.029104Z","shell.execute_reply.started":"2024-02-23T03:03:24.009245Z","shell.execute_reply":"2024-02-23T03:03:24.028359Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"case_ids = data[\"case_id\"].unique().shuffle(seed=1)\ncase_ids_train, case_ids_test = train_test_split(case_ids, train_size=0.6, random_state=1)\ncase_ids_valid, case_ids_test = train_test_split(case_ids_test, train_size=0.5, random_state=1)\n\ncols_pred = []\nfor col in data.columns:\n    if col[-1].isupper() and col[:-1].islower():\n        cols_pred.append(col)\n\nprint(cols_pred)\n\ndef from_polars_to_pandas(case_ids: pl.DataFrame) -> pl.DataFrame:\n    return (\n        data.filter(pl.col(\"case_id\").is_in(case_ids))[[\"case_id\", \"WEEK_NUM\", \"target\"]].to_pandas(),\n        data.filter(pl.col(\"case_id\").is_in(case_ids))[cols_pred].to_pandas(),\n        data.filter(pl.col(\"case_id\").is_in(case_ids))[\"target\"].to_pandas()\n    )\n\nbase_train, X_train, y_train = from_polars_to_pandas(case_ids_train)\nbase_valid, X_valid, y_valid = from_polars_to_pandas(case_ids_valid)\nbase_test, X_test, y_test = from_polars_to_pandas(case_ids_test)\n\nfor df in [X_train, X_valid, X_test]:\n    df = convert_strings(df)\n    \nprint('\\n')\nprint(f\"Train: {X_train.shape}\")\nprint(f\"Valid: {X_valid.shape}\")\nprint(f\"Test: {X_test.shape}\")","metadata":{"execution":{"iopub.status.busy":"2024-02-23T03:03:24.031817Z","iopub.execute_input":"2024-02-23T03:03:24.032409Z","iopub.status.idle":"2024-02-23T03:03:47.008292Z","shell.execute_reply.started":"2024-02-23T03:03:24.032383Z","shell.execute_reply":"2024-02-23T03:03:47.007377Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Training LightGBM\n\nMinimal example of LightGBM training is shown below.","metadata":{}},{"cell_type":"code","source":"# lgb_train = lgb.Dataset(X_train, label=y_train)\n# lgb_valid = lgb.Dataset(X_valid, label=y_valid, reference=lgb_train)\n\n# params = {\n#     \"boosting_type\": \"gbdt\",\n#     \"objective\": \"binary\",\n#     \"metric\": \"auc\",\n#     \"max_depth\": 3,\n#     \"num_leaves\": 31,\n#     \"learning_rate\": 0.05,\n#     \"feature_fraction\": 0.9,\n#     \"bagging_fraction\": 0.8,\n#     \"bagging_freq\": 5,\n#     \"n_estimators\": 1000,\n#     \"verbose\": -1,\n# }\n\n# gbm = lgb.train(\n#     params,\n#     lgb_train,\n#     valid_sets=lgb_valid,\n#     callbacks=[lgb.log_evaluation(50), lgb.early_stopping(10)]\n# )","metadata":{"execution":{"iopub.status.busy":"2024-02-21T01:57:44.876909Z","iopub.execute_input":"2024-02-21T01:57:44.877864Z","iopub.status.idle":"2024-02-21T01:59:56.610228Z","shell.execute_reply.started":"2024-02-21T01:57:44.877822Z","shell.execute_reply":"2024-02-21T01:59:56.607510Z"}},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# StratifiedGroupKFold\ncv = StratifiedGroupKFold(n_splits=5, shuffle=True, random_state=42)\n\nparams = {\n    \"boosting_type\": \"gbdt\",\n    \"objective\": \"binary\",\n    \"metric\": \"auc\",\n    \"max_depth\": 3,\n    \"num_leaves\": 31,\n    \"learning_rate\": 0.05,\n    \"feature_fraction\": 0.9,\n    \"bagging_fraction\": 0.8,\n    \"bagging_freq\": 5,\n    \"n_estimators\": 1000,\n    \"verbose\": -1,\n}\n\ncv_scores = []\n\nfor train_idx, valid_idx in cv.split(X_train, y_train, groups=base_train[\"WEEK_NUM\"]):\n    X_train_cv, X_valid_cv = X_train.iloc[train_idx], X_train.iloc[valid_idx]\n    y_train_cv, y_valid_cv = y_train.iloc[train_idx], y_train.iloc[valid_idx]\n\n    lgb_train = lgb.Dataset(X_train_cv, label=y_train_cv)\n    lgb_valid = lgb.Dataset(X_valid_cv, label=y_valid_cv, reference=lgb_train)\n\n    gbm = lgb.train(\n        params,\n        lgb_train,\n        valid_sets=lgb_valid,\n        callbacks=[lgb.log_evaluation(50), lgb.early_stopping(10)]\n    )\n\n    # 검증 데이터에 대한 예측\n    y_pred_valid = gbm.predict(X_valid_cv, num_iteration=gbm.best_iteration)\n    # AUC 스코어 계산\n    auc_score = roc_auc_score(y_valid_cv, y_pred_valid)\n    cv_scores.append(auc_score)\n\nprint(\"CV AUC scores: \", cv_scores)\nprint(\"Average CV AUC score: \", sum(cv_scores) / len(cv_scores))","metadata":{"execution":{"iopub.status.busy":"2024-02-23T03:06:02.901101Z","iopub.execute_input":"2024-02-23T03:06:02.902079Z","iopub.status.idle":"2024-02-23T03:12:55.585986Z","shell.execute_reply.started":"2024-02-23T03:06:02.902040Z","shell.execute_reply":"2024-02-23T03:12:55.584996Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Evaluation with AUC and then comparison with the stability metric is shown below.","metadata":{}},{"cell_type":"code","source":"# for base, X in [(base_train, X_train), (base_valid, X_valid), (base_test, X_test)]:\n#     y_pred = gbm.predict(X, num_iteration=gbm.best_iteration)\n#     base[\"score\"] = y_pred\n\n# print(f'The AUC score on the train set is: {roc_auc_score(base_train[\"target\"], base_train[\"score\"])}') \n# print(f'The AUC score on the valid set is: {roc_auc_score(base_valid[\"target\"], base_valid[\"score\"])}') \n# print(f'The AUC score on the test set is: {roc_auc_score(base_test[\"target\"], base_test[\"score\"])}')  \n\n\n# def gini_stability(base, w_fallingrate=88.0, w_resstd=-0.5):\n#     gini_in_time = base.loc[:, [\"WEEK_NUM\", \"target\", \"score\"]]\\\n#         .sort_values(\"WEEK_NUM\")\\\n#         .groupby(\"WEEK_NUM\")[[\"target\", \"score\"]]\\\n#         .apply(lambda x: 2*roc_auc_score(x[\"target\"], x[\"score\"])-1).tolist()\n    \n#     x = np.arange(len(gini_in_time))\n#     y = gini_in_time\n#     a, b = np.polyfit(x, y, 1)\n#     y_hat = a*x + b\n#     residuals = y - y_hat\n#     res_std = np.std(residuals)\n#     avg_gini = np.mean(gini_in_time)\n#     return avg_gini + w_fallingrate * min(0, a) + w_resstd * res_std\n\n# stability_score_train = gini_stability(base_train)\n# stability_score_valid = gini_stability(base_valid)\n# stability_score_test = gini_stability(base_test)\n\n# print('\\n')\n# print(f'The stability score on the train set is: {stability_score_train}') \n# print(f'The stability score on the valid set is: {stability_score_valid}') \n# print(f'The stability score on the test set is: {stability_score_test}')","metadata":{"execution":{"iopub.status.busy":"2024-02-21T01:59:56.614027Z","iopub.execute_input":"2024-02-21T01:59:56.616678Z","iopub.status.idle":"2024-02-21T02:00:33.152330Z","shell.execute_reply.started":"2024-02-21T01:59:56.616548Z","shell.execute_reply":"2024-02-21T02:00:33.151218Z"}},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Submission\n\nScoring the submission dataset is below, we need to take care of new categories. Then we save the score as a last step. ","metadata":{}},{"cell_type":"code","source":"# # 데이터 타입 일치 여부 확인\n# dtype_match = X_train.dtypes == X_submission.dtypes\n\n# if dtype_match.all():\n#     print(\"데이터 타입이 일치합니다.\")\n# else:\n#     print(\"데이터 타입이 일치하지 않습니다.\")\n    \n#     # 불일치한 열과 데이터 타입 출력\n#     mismatched_cols = X_train.columns[~dtype_match].tolist()\n#     for col in mismatched_cols:\n#         print(f\"데이터 타입 불일치 - 열: {col}, 훈련 데이터셋: {X_train[col].dtype}, 제출 데이터셋: {X_submission[col].dtype}\")","metadata":{"execution":{"iopub.status.busy":"2024-02-21T02:00:37.759178Z","iopub.execute_input":"2024-02-21T02:00:37.759684Z","iopub.status.idle":"2024-02-21T02:00:37.768881Z","shell.execute_reply.started":"2024-02-21T02:00:37.759644Z","shell.execute_reply":"2024-02-21T02:00:37.767797Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"X_submission = data_submission[cols_pred].to_pandas()\nX_submission = convert_strings(X_submission)\ncategorical_cols = X_train.select_dtypes(include=['category']).columns\n\nfor col in categorical_cols:\n    train_categories = set(X_train[col].cat.categories)\n    submission_categories = set(X_submission[col].cat.categories)\n    new_categories = submission_categories - train_categories\n    X_submission.loc[X_submission[col].isin(new_categories), col] = \"Unknown\"\n    new_dtype = pd.CategoricalDtype(categories=train_categories, ordered=True)\n    X_train[col] = X_train[col].astype(new_dtype)\n    X_submission[col] = X_submission[col].astype(new_dtype)\n\ny_submission_pred = gbm.predict(X_submission, num_iteration=gbm.best_iteration)\n\n\nsubmission = pd.DataFrame({\n    \"case_id\": data_submission[\"case_id\"].to_numpy(),\n    \"score\": y_submission_pred\n}).set_index('case_id')\nsubmission.to_csv(\"./submission.csv\")","metadata":{"execution":{"iopub.status.busy":"2024-02-23T03:14:32.506629Z","iopub.execute_input":"2024-02-23T03:14:32.507043Z","iopub.status.idle":"2024-02-23T03:14:33.033772Z","shell.execute_reply.started":"2024-02-23T03:14:32.507011Z","shell.execute_reply":"2024-02-23T03:14:33.032925Z"},"trusted":true},"execution_count":null,"outputs":[]}]}