{"metadata":{"kaggle":{"accelerator":"none","dataSources":[{"sourceId":50160,"databundleVersionId":7602123,"sourceType":"competition"}],"dockerImageVersionId":30648,"isInternetEnabled":false,"language":"python","sourceType":"notebook","isGpuEnabled":false},"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":4884.021459,"end_time":"2024-02-06T08:52:56.977348","environment_variables":{},"exception":null,"input_path":"__notebook__.ipynb","output_path":"__notebook__.ipynb","parameters":{},"start_time":"2024-02-06T07:31:32.955889","version":"2.4.0"}},"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\ndataPath = \"/kaggle/input/home-credit-credit-risk-model-stability/\"","metadata":{"papermill":{"duration":4.235504,"end_time":"2024-02-06T07:31:41.107510","exception":false,"start_time":"2024-02-06T07:31:36.872006","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-02-07T06:12:39.219898Z","iopub.execute_input":"2024-02-07T06:12:39.221078Z","iopub.status.idle":"2024-02-07T06:12:43.798647Z","shell.execute_reply.started":"2024-02-07T06:12:39.221038Z","shell.execute_reply":"2024-02-07T06:12:43.797481Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"This notebook is based on the work of Mr. Daniel Herman! Thanks for your amazing started. Here is the link https://www.kaggle.com/code/jetakow/home-credit-2024-starter-notebook.\n\nIn this notebook, I will present the cstboost model with 5 folds cross-validation. ","metadata":{}},{"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":{"papermill":{"duration":0.01679,"end_time":"2024-02-06T07:31:41.129274","exception":false,"start_time":"2024-02-06T07:31:41.112484","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-02-07T06:12:43.800948Z","iopub.execute_input":"2024-02-07T06:12:43.801471Z","iopub.status.idle":"2024-02-07T06:12:43.810799Z","shell.execute_reply.started":"2024-02-07T06:12:43.801429Z","shell.execute_reply":"2024-02-07T06:12:43.809778Z"},"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":{"papermill":{"duration":12.392163,"end_time":"2024-02-06T07:31:53.526418","exception":false,"start_time":"2024-02-06T07:31:41.134255","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-02-07T06:12:43.812159Z","iopub.execute_input":"2024-02-07T06:12:43.812504Z","iopub.status.idle":"2024-02-07T06:13:01.086286Z","shell.execute_reply.started":"2024-02-07T06:12:43.812479Z","shell.execute_reply":"2024-02-07T06:13:01.085365Z"},"trusted":true},"execution_count":null,"outputs":[]},{"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":{"papermill":{"duration":0.117972,"end_time":"2024-02-06T07:31:53.649206","exception":false,"start_time":"2024-02-06T07:31:53.531234","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-02-07T06:13:01.091376Z","iopub.execute_input":"2024-02-07T06:13:01.092180Z","iopub.status.idle":"2024-02-07T06:13:01.171748Z","shell.execute_reply.started":"2024-02-07T06:13:01.092138Z","shell.execute_reply":"2024-02-07T06:13:01.170629Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"## Feature engineering\n\nIn this part, we can see a simple example of joining tables via `case_id`. Here the loading and joining is done with polars library. Polars library is blazingly fast and has much smaller memory footprint than pandas. ","metadata":{"papermill":{"duration":0.004244,"end_time":"2024-02-06T07:31:53.658360","exception":false,"start_time":"2024-02-06T07:31:53.654116","status":"completed"},"tags":[]}},{"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":{"papermill":{"duration":1.263082,"end_time":"2024-02-06T07:31:54.925894","exception":false,"start_time":"2024-02-06T07:31:53.662812","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-02-07T06:13:01.173444Z","iopub.execute_input":"2024-02-07T06:13:01.173895Z","iopub.status.idle":"2024-02-07T06:13:04.100823Z","shell.execute_reply.started":"2024-02-07T06:13:01.173838Z","shell.execute_reply":"2024-02-07T06:13:04.099893Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"test_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\ntest_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\ntest_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\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":{"papermill":{"duration":0.426199,"end_time":"2024-02-06T07:31:55.356927","exception":false,"start_time":"2024-02-06T07:31:54.930728","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-02-07T06:13:04.101742Z","iopub.execute_input":"2024-02-07T06:13:04.102077Z","iopub.status.idle":"2024-02-07T06:13:05.640058Z","shell.execute_reply.started":"2024-02-07T06:13:04.102049Z","shell.execute_reply":"2024-02-07T06:13:05.639085Z"},"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":{"papermill":{"duration":4.984828,"end_time":"2024-02-06T07:32:00.346668","exception":false,"start_time":"2024-02-06T07:31:55.361840","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-02-07T06:13:05.640941Z","iopub.execute_input":"2024-02-07T06:13:05.641330Z","iopub.status.idle":"2024-02-07T06:13:12.795868Z","shell.execute_reply.started":"2024-02-07T06:13:05.641295Z","shell.execute_reply":"2024-02-07T06:13:12.794616Z"},"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":{"papermill":{"duration":0.013708,"end_time":"2024-02-06T07:32:00.365522","exception":false,"start_time":"2024-02-06T07:32:00.351814","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-02-07T06:13:12.797112Z","iopub.execute_input":"2024-02-07T06:13:12.797591Z","iopub.status.idle":"2024-02-07T06:13:12.804208Z","shell.execute_reply.started":"2024-02-07T06:13:12.797556Z","shell.execute_reply":"2024-02-07T06:13:12.802942Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# **Training**","metadata":{"papermill":{"duration":0.004343,"end_time":"2024-02-06T07:32:00.374801","exception":false,"start_time":"2024-02-06T07:32:00.370458","status":"completed"},"tags":[]}},{"cell_type":"code","source":"from sklearn.model_selection import StratifiedKFold\nfrom catboost import CatBoostClassifier\n\ncat_features = [col for col in X_train.columns if X_train[col].dtype.name == 'category' or X_train[col].dtype.name == 'object']\n\nfor col in cat_features:\n    X_train[col] = X_train[col].cat.add_categories('Missing').fillna('Missing')\n    X_valid[col] = X_valid[col].cat.add_categories('Missing').fillna('Missing')\n                                                                     \nparams = {\n    \"iterations\": 1500,\n    \"depth\": 15,\n    \"learning_rate\": 0.012473468,\n    \"eval_metric\": 'AUC',\n    \"random_seed\": 42,\n    \"bootstrap_type\": 'Bayesian',\n    \"bagging_temperature\": 1,\n    \"od_type\": 'Iter',\n    \"od_wait\": 50,\n    'l2_leaf_reg':0.234343736\n}\n\nskf = StratifiedKFold(n_splits=5, shuffle=True, random_state=31415926)\n\nauc_scores = []\nmodels = []\n\nfor train_index, valid_index in skf.split(X_train, y_train):\n    X_train_fold, X_valid_fold = X_train.iloc[train_index], X_train.iloc[valid_index]\n    y_train_fold, y_valid_fold = y_train.iloc[train_index], y_train.iloc[valid_index]\n\n    cat_model = CatBoostClassifier(**params)\n    cat_model.fit(\n        X_train_fold, y_train_fold,\n        eval_set=(X_valid_fold, y_valid_fold),\n        cat_features=cat_features,\n        use_best_model=True,\n        verbose=True\n    )\n    \n    models.append(cat_model)\n    \n    auc = cat_model.get_best_score()['validation']['AUC']\n    auc_scores.append(auc)\n\n# 计算平均 AUC 分数\naverage_auc = np.mean(auc_scores)\nprint(f\"Average AUC Score: {average_auc}\")\n","metadata":{"papermill":{"duration":2283.103416,"end_time":"2024-02-06T08:52:54.076015","exception":false,"start_time":"2024-02-06T08:14:50.972599","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-02-07T06:13:12.805946Z","iopub.execute_input":"2024-02-07T06:13:12.806254Z","iopub.status.idle":"2024-02-07T06:15:57.094202Z","shell.execute_reply.started":"2024-02-07T06:13:12.806230Z","shell.execute_reply":"2024-02-07T06:15:57.092691Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Evaluation with AUC and then comparison with the stability metric is shown below.","metadata":{"papermill":{"duration":0.064039,"end_time":"2024-02-06T08:52:54.203289","exception":false,"start_time":"2024-02-06T08:52:54.139250","status":"completed"},"tags":[]}},{"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":{"papermill":{"duration":0.062845,"end_time":"2024-02-06T08:52:54.607523","exception":false,"start_time":"2024-02-06T08:52:54.544678","status":"completed"},"tags":[]}},{"cell_type":"code","source":"X_submission = data_submission[cols_pred].to_pandas()\nX_submission = convert_strings(X_submission)\nX_submission_processed = X_submission.copy()\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\nfor col in cat_features:\n    X_submission_processed[col] = X_submission_processed[col].cat.add_categories('Missing').fillna('Missing')\n\n\npredictions = [model.predict_proba(X_submission_processed)[:,1] for model in models]","metadata":{"papermill":{"duration":0.204329,"end_time":"2024-02-06T08:52:54.873928","exception":false,"start_time":"2024-02-06T08:52:54.669599","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-02-07T06:15:57.097303Z","iopub.execute_input":"2024-02-07T06:15:57.097641Z","iopub.status.idle":"2024-02-07T06:15:57.225173Z","shell.execute_reply.started":"2024-02-07T06:15:57.097614Z","shell.execute_reply":"2024-02-07T06:15:57.223938Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"\ny_submission_pred = np.mean(predictions, axis=0)","metadata":{"execution":{"iopub.status.busy":"2024-02-07T06:15:57.226536Z","iopub.execute_input":"2024-02-07T06:15:57.227008Z","iopub.status.idle":"2024-02-07T06:15:57.233454Z","shell.execute_reply.started":"2024-02-07T06:15:57.226971Z","shell.execute_reply":"2024-02-07T06:15:57.232141Z"},"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":{"papermill":{"duration":0.089784,"end_time":"2024-02-06T08:52:55.026753","exception":false,"start_time":"2024-02-06T08:52:54.936969","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-02-07T06:15:57.235157Z","iopub.execute_input":"2024-02-07T06:15:57.236786Z","iopub.status.idle":"2024-02-07T06:15:57.255071Z","shell.execute_reply.started":"2024-02-07T06:15:57.236738Z","shell.execute_reply":"2024-02-07T06:15:57.253672Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"submission","metadata":{"papermill":{"duration":0.086803,"end_time":"2024-02-06T08:52:55.176296","exception":false,"start_time":"2024-02-06T08:52:55.089493","status":"completed"},"tags":[],"execution":{"iopub.status.busy":"2024-02-07T06:15:57.256577Z","iopub.execute_input":"2024-02-07T06:15:57.256938Z","iopub.status.idle":"2024-02-07T06:15:57.274301Z","shell.execute_reply.started":"2024-02-07T06:15:57.256905Z","shell.execute_reply":"2024-02-07T06:15:57.272907Z"},"trusted":true},"execution_count":null,"outputs":[]}]}