{"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":"gpu","dataSources":[{"sourceId":50160,"databundleVersionId":7921029,"sourceType":"competition"}],"dockerImageVersionId":30635,"isInternetEnabled":false,"language":"python","sourceType":"notebook","isGpuEnabled":true}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"Я взял бейзлайновый ноутбук от автора чтобы получить сразу рабочий датасет\n\nГде коменты на русском мои содержательные новвоведения","metadata":{}},{"cell_type":"markdown","source":"## Load the data","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19"}},{"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":"2024-03-16T10:48:17.910853Z","iopub.execute_input":"2024-03-16T10:48:17.911356Z","iopub.status.idle":"2024-03-16T10:48:23.573494Z","shell.execute_reply.started":"2024-03-16T10:48:17.911321Z","shell.execute_reply":"2024-03-16T10:48:23.572717Z"},"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":"2024-03-16T10:48:23.575482Z","iopub.execute_input":"2024-03-16T10:48:23.575815Z","iopub.status.idle":"2024-03-16T10:48:23.584723Z","shell.execute_reply.started":"2024-03-16T10:48:23.575782Z","shell.execute_reply":"2024-03-16T10:48:23.583543Z"},"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":"2024-03-16T10:48:23.585715Z","iopub.execute_input":"2024-03-16T10:48:23.585969Z","iopub.status.idle":"2024-03-16T10:48:37.856175Z","shell.execute_reply.started":"2024-03-16T10:48:23.585945Z","shell.execute_reply":"2024-03-16T10:48:37.855194Z"},"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":{"execution":{"iopub.status.busy":"2024-03-16T10:48:37.857388Z","iopub.execute_input":"2024-03-16T10:48:37.857680Z","iopub.status.idle":"2024-03-16T10:48:37.907497Z","shell.execute_reply.started":"2024-03-16T10:48:37.857655Z","shell.execute_reply":"2024-03-16T10:48:37.906796Z"},"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":{}},{"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":"2024-03-16T10:48:37.910693Z","iopub.execute_input":"2024-03-16T10:48:37.911339Z","iopub.status.idle":"2024-03-16T10:48:39.318157Z","shell.execute_reply.started":"2024-03-16T10:48:37.911305Z","shell.execute_reply":"2024-03-16T10:48:39.317138Z"},"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":"2024-03-16T10:48:39.319273Z","iopub.execute_input":"2024-03-16T10:48:39.319553Z","iopub.status.idle":"2024-03-16T10:48:39.330574Z","shell.execute_reply.started":"2024-03-16T10:48:39.319529Z","shell.execute_reply":"2024-03-16T10:48:39.329694Z"},"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":"2024-03-16T10:48:39.331643Z","iopub.execute_input":"2024-03-16T10:48:39.331955Z","iopub.status.idle":"2024-03-16T10:48:46.746562Z","shell.execute_reply.started":"2024-03-16T10:48:39.331917Z","shell.execute_reply":"2024-03-16T10:48:46.745587Z"},"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":"2024-03-16T10:48:46.747791Z","iopub.execute_input":"2024-03-16T10:48:46.748181Z","iopub.status.idle":"2024-03-16T10:48:46.753574Z","shell.execute_reply.started":"2024-03-16T10:48:46.748130Z","shell.execute_reply":"2024-03-16T10:48:46.752590Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Тут заканчивается базовая обработка датасета и начинается построение модели","metadata":{}},{"cell_type":"code","source":"X_train.columns","metadata":{"execution":{"iopub.status.busy":"2024-03-16T10:48:46.754729Z","iopub.execute_input":"2024-03-16T10:48:46.755009Z","iopub.status.idle":"2024-03-16T10:48:46.767657Z","shell.execute_reply.started":"2024-03-16T10:48:46.754985Z","shell.execute_reply":"2024-03-16T10:48:46.766846Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"Cоздание фич","metadata":{}},{"cell_type":"code","source":"# Определение функции для создания новых признаков\ndef create_features(df):\n    # Преобразование суммы аннуитета в логарифмический масштаб\n    df['log_annuity_780A'] = np.log1p(df['annuity_780A'])\n    \n    # Соотношение средних просмотров к среднему количеству платежей\n    df['ratio_avgviews_avgpayments'] = df['avgpmtlast12m_4525200A'] / df['avginstallast24m_3658937A']\n    \n    # Соотношение среднего количества платежей к общей сумме долга\n    df['ratio_avgpayments_totaldebt'] = df['avgpmtlast12m_4525200A'] / df['totaldebt_9A']\n\n    # Стандартное отклонение по определенным платежным колонкам\n    df['std_payments'] = df[['avgpmtlast12m_4525200A', 'maxpmtlast3m_4525190A']].std(axis=1)\n    \n    # Количество кредитов, сумма которых выше среднего значения кредита\n    mean_cred_amount = df['credamount_770A'].mean()\n    df['count_credamount_above_mean'] = (df['credamount_770A'] > mean_cred_amount).astype(int)\n    \n    return df\n\n# Применение функции к каждому DataFrame\nX_train = create_features(X_train)\nX_valid = create_features(X_valid)\nX_test = create_features(X_test)\n","metadata":{"execution":{"iopub.status.busy":"2024-03-16T10:48:46.768773Z","iopub.execute_input":"2024-03-16T10:48:46.769122Z","iopub.status.idle":"2024-03-16T10:48:47.312974Z","shell.execute_reply.started":"2024-03-16T10:48:46.769090Z","shell.execute_reply":"2024-03-16T10:48:47.312172Z"},"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":"from sklearn.model_selection import RandomizedSearchCV\nimport lightgbm as lgb\n\n# Обновлённые значения гиперпараметров с вариациями для max_depth и learning_rate\nparam_dist = {\n    'boosting_type': ['gbdt'],  # без изменений\n    'objective': ['binary'],    # без изменений\n    'metric': ['auc'],          # без изменений\n    'max_depth': [3, 5],        # теперь два варианта: 3 или 5\n    'num_leaves': [31, 50],     # добавлен вариант 50 к уже существующему 31\n    'learning_rate': [0.05, 0.01],  # теперь два варианта: 0.05 или 0.01\n    'feature_fraction': [0.9],  # без изменений\n    'bagging_fraction': [0.8],  # без изменений\n    'bagging_freq': [5],        # без изменений\n    'n_estimators': [1000],     # без изменений\n    'verbose': [-1],            # без изменений\n}\n\n# Создайте и обучите модель LightGBM с перебором гиперпараметров\nlgb_estimator = lgb.LGBMClassifier(random_state=42)\n\nrandom_search = RandomizedSearchCV(\n    estimator=lgb_estimator,\n    param_distributions=param_dist,\n    n_iter=10,  # Увеличено количество итераций для исследования различных комбинаций\n    scoring='roc_auc',\n    cv=3,\n    verbose=1,\n    random_state=42\n)\n\n# Обучаем с перебором гиперпараметров\nrandom_search.fit(X_train, y_train)\n\n# Вывод лучших параметров и их оценки\nprint(\"Лучшие параметры:\", random_search.best_params_)\nprint(\"Лучшая оценка AUC:\", random_search.best_score_)\n\n# Модель с лучшими параметрами\nbest_model = random_search.best_estimator_\n","metadata":{"execution":{"iopub.status.busy":"2024-03-16T10:48:47.314193Z","iopub.execute_input":"2024-03-16T10:48:47.314526Z","iopub.status.idle":"2024-03-16T10:52:59.503495Z","shell.execute_reply.started":"2024-03-16T10:48:47.314497Z","shell.execute_reply":"2024-03-16T10:52:59.502677Z"},"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 = best_model.predict_proba(X)[:, 1]  # Используем predict_proba для получения вероятностей\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\"])}') \n","metadata":{"execution":{"iopub.status.busy":"2024-03-16T10:52:59.506807Z","iopub.execute_input":"2024-03-16T10:52:59.509123Z","iopub.status.idle":"2024-03-16T10:53:30.556958Z","shell.execute_reply.started":"2024-03-16T10:52:59.509090Z","shell.execute_reply":"2024-03-16T10:53:30.556044Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"было\n\nThe AUC score on the train set is: 0.764122917660593\n\nThe AUC score on the valid set is: 0.7512157223309048\n\nThe AUC score on the test set is: 0.7483072129459662\n\nвидно что скор улучшился","metadata":{}},{"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":"2024-03-16T10:53:30.558242Z","iopub.execute_input":"2024-03-16T10:53:30.558555Z","iopub.status.idle":"2024-03-16T10:53:31.531106Z","shell.execute_reply.started":"2024-03-16T10:53:30.558529Z","shell.execute_reply":"2024-03-16T10:53:31.530177Z"},"trusted":true},"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":"X_submission = data_submission[cols_pred].to_pandas()\nX_submission = convert_strings(X_submission)\ncategorical_cols = X_train.select_dtypes(include=['category']).columns\n\n# Применяем те же преобразования категориальных признаков\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_submission[col] = X_submission[col].astype(new_dtype)\n\n    \nX_submission = create_features(X_submission)\n# Применяем лучшую модель для получения предсказаний\ny_submission_pred = best_model.predict_proba(X_submission)[:, 1]","metadata":{"execution":{"iopub.status.busy":"2024-03-16T10:53:31.533904Z","iopub.execute_input":"2024-03-16T10:53:31.534207Z","iopub.status.idle":"2024-03-16T10:53:31.594918Z","shell.execute_reply.started":"2024-03-16T10:53:31.534182Z","shell.execute_reply":"2024-03-16T10:53:31.594196Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# Сохраняем предсказания\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":"2024-03-16T10:53:31.595835Z","iopub.execute_input":"2024-03-16T10:53:31.596266Z","iopub.status.idle":"2024-03-16T10:53:31.606631Z","shell.execute_reply.started":"2024-03-16T10:53:31.596239Z","shell.execute_reply":"2024-03-16T10:53:31.605921Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"","metadata":{},"execution_count":null,"outputs":[]}]}