{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","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"},"kaggle":{"accelerator":"none","dataSources":[{"sourceId":50160,"databundleVersionId":7921029,"sourceType":"competition"}],"dockerImageVersionId":30664,"isInternetEnabled":false,"language":"python","sourceType":"notebook","isGpuEnabled":false}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"# Imports","metadata":{}},{"cell_type":"markdown","source":"Matvey Ryumin on the base of: https://www.kaggle.com/code/rishabh15virgo/home-credit-data-understanding-baseline\nI have tuned the hyperparameters through ml-package optuna. Then trained the model with the best hyperparameters and obtained the results. See the Hyperparameters search\n\n\n- Initial public score: 0.360\n- The AUC score on the train set is: 0.764122917660593\n- The AUC score on the valid set is: 0.7512157223309048\n- The AUC score on the test set is: 0.7483072129459662\n- The stability score on the train set is: 0.4976648127691175\n- The stability score on the valid set is: 0.4726726686264489\n- The stability score on the test set is: 0.4583643686935092\n\nNew values:\n- The AUC score on the train set is: 0.8080183072413502\n- The AUC score on the valid set is: 0.7473949778991889\n- The AUC score on the test set is: 0.7447609514712573\n- The stability score on the train set is: 0.48000221220191697\n- The stability score on the valid set is: 0.4623273808370703\n- The stability score on the test set is: 0.44936835939499","metadata":{}},{"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\ndataPath = \"/kaggle/input/home-credit-credit-risk-model-stability/\"","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:26:28.235825Z","iopub.execute_input":"2024-03-17T18:26:28.236364Z","iopub.status.idle":"2024-03-17T18:26:28.248010Z","shell.execute_reply.started":"2024-03-17T18:26:28.236331Z","shell.execute_reply":"2024-03-17T18:26:28.246141Z"},"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\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":"2024-03-17T18:26:28.250446Z","iopub.execute_input":"2024-03-17T18:26:28.251441Z","iopub.status.idle":"2024-03-17T18:26:28.261891Z","shell.execute_reply.started":"2024-03-17T18:26:28.251408Z","shell.execute_reply":"2024-03-17T18:26:28.260708Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_basetable = pl.read_csv(dataPath + \"csv_files/train/train_base.csv\")\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)\ntrain_person_1 = pl.read_csv(dataPath + \"csv_files/train/train_person_1.csv\").pipe(set_table_dtypes) \ntrain_credit_bureau_b_2 = pl.read_csv(dataPath + \"csv_files/train/train_credit_bureau_b_2.csv\").pipe(set_table_dtypes) ","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:26:28.263467Z","iopub.execute_input":"2024-03-17T18:26:28.264481Z","iopub.status.idle":"2024-03-17T18:26:53.355391Z","shell.execute_reply.started":"2024-03-17T18:26:28.264434Z","shell.execute_reply":"2024-03-17T18:26:53.352639Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### basetable contains decision dates and target","metadata":{}},{"cell_type":"code","source":"train_basetable.shape","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:26:53.361617Z","iopub.execute_input":"2024-03-17T18:26:53.362315Z","iopub.status.idle":"2024-03-17T18:26:53.373996Z","shell.execute_reply.started":"2024-03-17T18:26:53.362268Z","shell.execute_reply":"2024-03-17T18:26:53.372559Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_basetable.head()","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:26:53.375423Z","iopub.execute_input":"2024-03-17T18:26:53.375863Z","iopub.status.idle":"2024-03-17T18:26:53.395464Z","shell.execute_reply.started":"2024-03-17T18:26:53.375836Z","shell.execute_reply":"2024-03-17T18:26:53.394146Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### static table contains paid related features","metadata":{}},{"cell_type":"code","source":"train_static.shape","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:26:53.396425Z","iopub.execute_input":"2024-03-17T18:26:53.396741Z","iopub.status.idle":"2024-03-17T18:26:53.407576Z","shell.execute_reply.started":"2024-03-17T18:26:53.396713Z","shell.execute_reply":"2024-03-17T18:26:53.406221Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_static.head()","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:26:53.409312Z","iopub.execute_input":"2024-03-17T18:26:53.409620Z","iopub.status.idle":"2024-03-17T18:26:53.435252Z","shell.execute_reply.started":"2024-03-17T18:26:53.409595Z","shell.execute_reply":"2024-03-17T18:26:53.433911Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### Static_cb Contains personal information like DOB, education, marital status","metadata":{}},{"cell_type":"code","source":"train_static_cb.shape","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:26:53.437855Z","iopub.execute_input":"2024-03-17T18:26:53.438257Z","iopub.status.idle":"2024-03-17T18:26:53.446549Z","shell.execute_reply.started":"2024-03-17T18:26:53.438226Z","shell.execute_reply":"2024-03-17T18:26:53.445226Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_static_cb.head()","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:26:53.448753Z","iopub.execute_input":"2024-03-17T18:26:53.449309Z","iopub.status.idle":"2024-03-17T18:26:53.468083Z","shell.execute_reply.started":"2024-03-17T18:26:53.449277Z","shell.execute_reply":"2024-03-17T18:26:53.466561Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_static_cb[\"maritalst_385M\"].unique()","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:26:53.473420Z","iopub.execute_input":"2024-03-17T18:26:53.474125Z","iopub.status.idle":"2024-03-17T18:26:53.513178Z","shell.execute_reply.started":"2024-03-17T18:26:53.474091Z","shell.execute_reply":"2024-03-17T18:26:53.512045Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_static_cb[\"maritalst_893M\"].unique()","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:26:53.515062Z","iopub.execute_input":"2024-03-17T18:26:53.515426Z","iopub.status.idle":"2024-03-17T18:26:53.551835Z","shell.execute_reply.started":"2024-03-17T18:26:53.515396Z","shell.execute_reply":"2024-03-17T18:26:53.550684Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### Person personal info like gender, zipcode district etc","metadata":{}},{"cell_type":"code","source":"train_person_1.shape","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:26:53.553630Z","iopub.execute_input":"2024-03-17T18:26:53.553969Z","iopub.status.idle":"2024-03-17T18:26:53.561870Z","shell.execute_reply.started":"2024-03-17T18:26:53.553942Z","shell.execute_reply":"2024-03-17T18:26:53.560766Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_person_1.head()","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:26:53.563526Z","iopub.execute_input":"2024-03-17T18:26:53.563842Z","iopub.status.idle":"2024-03-17T18:26:53.583059Z","shell.execute_reply.started":"2024-03-17T18:26:53.563816Z","shell.execute_reply":"2024-03-17T18:26:53.582193Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### credit_bureau_b_2 ? need more information","metadata":{}},{"cell_type":"code","source":"train_credit_bureau_b_2.shape","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:26:53.585128Z","iopub.execute_input":"2024-03-17T18:26:53.585455Z","iopub.status.idle":"2024-03-17T18:26:53.598788Z","shell.execute_reply.started":"2024-03-17T18:26:53.585429Z","shell.execute_reply":"2024-03-17T18:26:53.597383Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train_credit_bureau_b_2.head()","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:26:53.600564Z","iopub.execute_input":"2024-03-17T18:26:53.600906Z","iopub.status.idle":"2024-03-17T18:26:53.614448Z","shell.execute_reply.started":"2024-03-17T18:26:53.600876Z","shell.execute_reply":"2024-03-17T18:26:53.613021Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Test data","metadata":{}},{"cell_type":"code","source":"test_basetable = pl.read_csv(dataPath + \"csv_files/test/test_base.csv\")\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)\ntest_person_1 = pl.read_csv(dataPath + \"csv_files/test/test_person_1.csv\").pipe(set_table_dtypes) \ntest_credit_bureau_b_2 = pl.read_csv(dataPath + \"csv_files/test/test_credit_bureau_b_2.csv\").pipe(set_table_dtypes) ","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:26:53.616933Z","iopub.execute_input":"2024-03-17T18:26:53.617421Z","iopub.status.idle":"2024-03-17T18:26:53.686524Z","shell.execute_reply.started":"2024-03-17T18:26:53.617380Z","shell.execute_reply":"2024-03-17T18:26:53.683383Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Feature engineering","metadata":{}},{"cell_type":"code","source":"# We need to use aggregation functions in tables with depth > 1, so tables that contain num_group1 column or \n# also num_group2 column.\ntrain_person_1_feats_1 = train_person_1.group_by(\"case_id\").agg(\n    pl.col(\"mainoccupationinc_384A\").max().alias(\"mainoccupationinc_384A_max\"),\n    (pl.col(\"incometype_1044T\") == \"SELFEMPLOYED\").max().alias(\"mainoccupationinc_384A_any_selfemployed\")\n)\n\n# Here num_group1=0 has special meaning, it is the person who applied for the loan.\ntrain_person_1_feats_2 = train_person_1.select([\"case_id\", \"num_group1\", \"housetype_905L\"]).filter(\n    pl.col(\"num_group1\") == 0\n).drop(\"num_group1\").rename({\"housetype_905L\": \"person_housetype\"})\n\n# Here we have num_goup1 and num_group2, so we need to aggregate again.\ntrain_credit_bureau_b_2_feats = train_credit_bureau_b_2.group_by(\"case_id\").agg(\n    pl.col(\"pmts_pmtsoverdue_635A\").max().alias(\"pmts_pmtsoverdue_635A_max\"),\n    (pl.col(\"pmts_dpdvalue_108P\") > 31).max().alias(\"pmts_dpdvalue_108P_over31\")\n)\n\n# We will process in this examples only A-type and M-type columns, so we need to select them.\nselected_static_cols = []\nfor col in train_static.columns:\n    if col[-1] in (\"A\", \"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\", \"M\"):\n        selected_static_cb_cols.append(col)\nprint(selected_static_cb_cols)\n\n# Join all tables together.\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_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_2_feats, how=\"left\", on=\"case_id\"\n)","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:26:53.688892Z","iopub.execute_input":"2024-03-17T18:26:53.689470Z","iopub.status.idle":"2024-03-17T18:27:01.117832Z","shell.execute_reply.started":"2024-03-17T18:26:53.689399Z","shell.execute_reply":"2024-03-17T18:27:01.116684Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"test_person_1_feats_1 = test_person_1.group_by(\"case_id\").agg(\n    pl.col(\"mainoccupationinc_384A\").max().alias(\"mainoccupationinc_384A_max\"),\n    (pl.col(\"incometype_1044T\") == \"SELFEMPLOYED\").max().alias(\"mainoccupationinc_384A_any_selfemployed\")\n)\n\ntest_person_1_feats_2 = test_person_1.select([\"case_id\", \"num_group1\", \"housetype_905L\"]).filter(\n    pl.col(\"num_group1\") == 0\n).drop(\"num_group1\").rename({\"housetype_905L\": \"person_housetype\"})\n\ntest_credit_bureau_b_2_feats = test_credit_bureau_b_2.group_by(\"case_id\").agg(\n    pl.col(\"pmts_pmtsoverdue_635A\").max().alias(\"pmts_pmtsoverdue_635A_max\"),\n    (pl.col(\"pmts_dpdvalue_108P\") > 31).max().alias(\"pmts_dpdvalue_108P_over31\")\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_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_2_feats, how=\"left\", on=\"case_id\"\n)","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:27:01.119643Z","iopub.execute_input":"2024-03-17T18:27:01.119976Z","iopub.status.idle":"2024-03-17T18:27:01.136207Z","shell.execute_reply.started":"2024-03-17T18:27:01.119950Z","shell.execute_reply":"2024-03-17T18:27:01.134958Z"},"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)","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:27:01.137803Z","iopub.execute_input":"2024-03-17T18:27:01.138188Z","iopub.status.idle":"2024-03-17T18:27:18.322755Z","shell.execute_reply.started":"2024-03-17T18:27:01.138144Z","shell.execute_reply":"2024-03-17T18:27:18.321275Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(f\"Train: {X_train.shape}\")\nprint(f\"Valid: {X_valid.shape}\")\nprint(f\"Test: {X_test.shape}\")","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:27:18.324999Z","iopub.execute_input":"2024-03-17T18:27:18.325867Z","iopub.status.idle":"2024-03-17T18:27:18.333493Z","shell.execute_reply.started":"2024-03-17T18:27:18.325808Z","shell.execute_reply":"2024-03-17T18:27:18.332370Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(\"Number of loan defaulters in train data : \",y_train.sum())\nprint(\"Percentage of loan defaulters in train data : \",(y_train.sum()/len(y_train)*100))\n\nprint(\"Number of loan defaulters in valid data : \",y_valid.sum())\nprint(\"Percentage of loan defaulters in valid data : \",(y_valid.sum()/len(y_valid)*100))\n\nprint(\"Number of loan defaulters in test data : \",y_test.sum())\nprint(\"Percentage of loan defaulters in test data : \",(y_test.sum()/len(y_test)*100))","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:27:18.334832Z","iopub.execute_input":"2024-03-17T18:27:18.336235Z","iopub.status.idle":"2024-03-17T18:27:18.352123Z","shell.execute_reply.started":"2024-03-17T18:27:18.336182Z","shell.execute_reply":"2024-03-17T18:27:18.350704Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Baseline model","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\": 2000,\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":"2024-03-17T18:27:18.353635Z","iopub.execute_input":"2024-03-17T18:27:18.354438Z","iopub.status.idle":"2024-03-17T18:27:18.360629Z","shell.execute_reply.started":"2024-03-17T18:27:18.354400Z","shell.execute_reply":"2024-03-17T18:27:18.359295Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# # Initial\n# 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\"])}')  ","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:27:18.362661Z","iopub.execute_input":"2024-03-17T18:27:18.363411Z","iopub.status.idle":"2024-03-17T18:27:18.377011Z","shell.execute_reply.started":"2024-03-17T18:27:18.363365Z","shell.execute_reply":"2024-03-17T18:27:18.375445Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# 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(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":"2024-03-17T18:27:18.378887Z","iopub.execute_input":"2024-03-17T18:27:18.379393Z","iopub.status.idle":"2024-03-17T18:27:18.387553Z","shell.execute_reply.started":"2024-03-17T18:27:18.379350Z","shell.execute_reply":"2024-03-17T18:27:18.386597Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Hyperparameters search","metadata":{}},{"cell_type":"code","source":"#!pip install optuna","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:27:18.388730Z","iopub.execute_input":"2024-03-17T18:27:18.389221Z","iopub.status.idle":"2024-03-17T18:27:18.403462Z","shell.execute_reply.started":"2024-03-17T18:27:18.389179Z","shell.execute_reply":"2024-03-17T18:27:18.402355Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import optuna\n\nlgb_train = lgb.Dataset(X_train, label=y_train)\nlgb_valid = lgb.Dataset(X_valid, label=y_valid, reference=lgb_train)\n\n\ndef objective(trial):\n    param = {\n        \"objective\": \"binary\",\n        \"metric\": \"auc\",\n        \"verbosity\": -1,\n        \"boosting_type\": \"gbdt\",\n        \"num_leaves\": trial.suggest_int(\"num_leaves\", 2, 128),\n        \"max_depth\": trial.suggest_int(\"max_depth\", -1, 32),\n        \"learning_rate\": trial.suggest_float(\"learning_rate\", 0.001, 0.3),\n        \"n_estimators\": trial.suggest_int(\"n_estimators\", 100, 1000),\n        \"feature_fraction\": trial.suggest_float(\"feature_fraction\", 0.4, 1.0),\n        \"bagging_fraction\": trial.suggest_float(\"bagging_fraction\", 0.4, 1.0),\n        \"bagging_freq\": trial.suggest_int(\"bagging_freq\", 1, 7)\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(5)])\n    \n    auc = gbm.best_score[\"valid_0\"][\"auc\"]\n    return auc\n\nstudy = optuna.create_study(direction=\"maximize\")\nstudy.optimize(objective, n_trials=50)\n\nprint(\"Best trial:\", study.best_trial.params)","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:27:18.405133Z","iopub.execute_input":"2024-03-17T18:27:18.406037Z","iopub.status.idle":"2024-03-17T18:27:18.415952Z","shell.execute_reply.started":"2024-03-17T18:27:18.405983Z","shell.execute_reply":"2024-03-17T18:27:18.414992Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(\"Best trial:\", study.best_trial.params)","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:27:18.417473Z","iopub.execute_input":"2024-03-17T18:27:18.418793Z","iopub.status.idle":"2024-03-17T18:27:18.430973Z","shell.execute_reply.started":"2024-03-17T18:27:18.418744Z","shell.execute_reply":"2024-03-17T18:27:18.430124Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import lightgbm as lgb\nfrom sklearn.metrics import roc_auc_score\n\n# Best parameters from the tuning process\nbest_params = {\n    \"boosting_type\": \"gbdt\",\n    \"objective\": \"binary\",\n    \"metric\": \"auc\",\n    \"num_leaves\": 88,\n    \"max_depth\": 15,\n    \"learning_rate\": 0.16472415456175532,\n    \"n_estimators\": 708,\n    \"feature_fraction\": 0.7654221581472358,\n    \"bagging_fraction\": 0.8383881018821167,\n    \"bagging_freq\": 3,\n    \"verbose\": -1\n}\n\n# Creating datasets for LightGBM\nlgb_train = lgb.Dataset(X_train, label=y_train)\nlgb_valid = lgb.Dataset(X_valid, label=y_valid, reference=lgb_train)\n\n# Train the model\ngbm = lgb.train(\nbest_params,\nlgb_train,\nvalid_sets=lgb_valid,\ncallbacks=[lgb.log_evaluation(50), lgb.early_stopping(10)])","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:38:54.467362Z","iopub.execute_input":"2024-03-17T18:38:54.467927Z","iopub.status.idle":"2024-03-17T18:39:13.452433Z","shell.execute_reply.started":"2024-03-17T18:38:54.467891Z","shell.execute_reply":"2024-03-17T18:39:13.451176Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# my model\nfor 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\nprint(f'The AUC score on the train set is: {roc_auc_score(base_train[\"target\"], base_train[\"score\"])}') \nprint(f'The AUC score on the valid set is: {roc_auc_score(base_valid[\"target\"], base_valid[\"score\"])}') \nprint(f'The AUC score on the test set is: {roc_auc_score(base_test[\"target\"], base_test[\"score\"])}')  ","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:41:28.381517Z","iopub.execute_input":"2024-03-17T18:41:28.381965Z","iopub.status.idle":"2024-03-17T18:41:32.859693Z","shell.execute_reply.started":"2024-03-17T18:41:28.381932Z","shell.execute_reply":"2024-03-17T18:41:32.858543Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"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\nstability_score_train = gini_stability(base_train)\nstability_score_valid = gini_stability(base_valid)\nstability_score_test = gini_stability(base_test)\n\nprint(f'The stability score on the train set is: {stability_score_train}') \nprint(f'The stability score on the valid set is: {stability_score_valid}') \nprint(f'The stability score on the test set is: {stability_score_test}')  ","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:28:27.982647Z","iopub.execute_input":"2024-03-17T18:28:27.983022Z","iopub.status.idle":"2024-03-17T18:28:29.065993Z","shell.execute_reply.started":"2024-03-17T18:28:27.982991Z","shell.execute_reply":"2024-03-17T18:28:29.064744Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Submission","metadata":{}},{"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)","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:28:29.067553Z","iopub.execute_input":"2024-03-17T18:28:29.068123Z","iopub.status.idle":"2024-03-17T18:28:29.180846Z","shell.execute_reply.started":"2024-03-17T18:28:29.068082Z","shell.execute_reply":"2024-03-17T18:28:29.179687Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"submission = 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":"2024-03-17T18:28:29.182321Z","iopub.execute_input":"2024-03-17T18:28:29.182750Z","iopub.status.idle":"2024-03-17T18:28:29.192719Z","shell.execute_reply.started":"2024-03-17T18:28:29.182711Z","shell.execute_reply":"2024-03-17T18:28:29.191550Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"submission.head()","metadata":{"execution":{"iopub.status.busy":"2024-03-17T18:28:29.194148Z","iopub.execute_input":"2024-03-17T18:28:29.194600Z","iopub.status.idle":"2024-03-17T18:28:29.209184Z","shell.execute_reply.started":"2024-03-17T18:28:29.194562Z","shell.execute_reply":"2024-03-17T18:28:29.207873Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]}]}