{"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":"gpu","dataSources":[{"sourceId":50160,"databundleVersionId":7921029,"sourceType":"competition"}],"dockerImageVersionId":30665,"isInternetEnabled":false,"language":"python","sourceType":"notebook","isGpuEnabled":true}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"import sys\nfrom pathlib import Path\nimport subprocess\nimport os\nimport gc\nfrom glob import glob\n\nimport numpy as np\nimport pandas as pd\nimport polars as pl\nfrom datetime import datetime\nimport seaborn as sns\nimport matplotlib.pyplot as plt\n\nimport warnings\nwarnings.filterwarnings('ignore')\n\nROOT = '/kaggle/input/home-credit-credit-risk-model-stability'\n\nfrom sklearn.model_selection import TimeSeriesSplit, GroupKFold, StratifiedGroupKFold\nfrom sklearn.base import BaseEstimator, RegressorMixin\nfrom sklearn.metrics import roc_auc_score\nimport lightgbm as lgb\n\nfrom imblearn.over_sampling import SMOTE\nfrom sklearn.preprocessing import OrdinalEncoder\nfrom sklearn.impute import KNNImputer","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","execution":{"iopub.status.busy":"2024-06-06T05:35:41.722527Z","iopub.execute_input":"2024-06-06T05:35:41.723143Z","iopub.status.idle":"2024-06-06T05:35:44.870260Z","shell.execute_reply.started":"2024-06-06T05:35:41.723109Z","shell.execute_reply":"2024-06-06T05:35:44.869363Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"class Pipeline:     \n    \n    # 进行数据预处理工作，对数据的格式进行转换\n\n    def set_table_dtypes(df):\n        for col in df.columns:\n            if col in [\"case_id\", \"WEEK_NUM\", \"num_group1\", \"num_group2\"]:\n                df = df.with_columns(pl.col(col).cast(pl.Int64))      # 将 case_id,WEEK_NUM，num_group1这种数据变为Int64类型的\n            elif col in [\"date_decision\"]:\n                df = df.with_columns(pl.col(col).cast(pl.Date))       # 将date_decision设置成date日期类型数据\n            elif col[-1] in (\"P\", \"A\"):\n                df = df.with_columns(pl.col(col).cast(pl.Float64))    # 将P，A打头的处理成float64类型\n            elif col[-1] in (\"M\",):    # Tan：尝试将str中部分进行Target Encoding\n                df = df.with_columns(pl.col(col).cast(pl.String))     # to字符串\n            elif col[-1] in (\"D\",):\n                df = df.with_columns(pl.col(col).cast(pl.Date))       # 处理成日期\n        return df\n\n    def handle_dates(df):                       #  对日期处理的函数，将日期处理成 截止日期减去开始日期的差值\n        for col in df.columns:\n            if col[-1] in (\"D\",):\n                df = df.with_columns(pl.col(col) - pl.col(\"date_decision\"))  #!!?\n                df = df.with_columns(pl.col(col).dt.total_days()) # t - t-1\n        df = df.drop(\"date_decision\", \"MONTH\")\n        return df\n\n    def filter_cols(df):                    # 去除掉数据缺失值达到70%的列，以及进行常数、高基数的去除\n        for col in df.columns:\n            if col not in [\"target\", \"case_id\", \"WEEK_NUM\"]:\n                isnull = df[col].is_null().mean()\n                if isnull > 0.7:\n                    df = df.drop(col)\n        \n        for col in df.columns:\n            if (col not in [\"target\", \"case_id\", \"WEEK_NUM\"]) & (df[col].dtype == pl.String):\n                freq = df[col].n_unique()\n                if (freq == 1) | (freq > 200):\n                    df = df.drop(col)\n        \n        return df\n\n\n    \nclass Aggregator:             # 用于从现有的DataFrame中生成用于机器学习的额外特征      生成额外特征\n    #Please add or subtract features yourself, be aware that too many features will take up too much space.\n    def num_expr(df):         \n         # 专注于以 \"P\" 或 \"A\" 结尾的列（可能是数值特征）。它生成最大值、最后值和平均值表达式。\n        cols = [col for col in df.columns if col[-1] in (\"P\", \"A\")]\n        expr_max = [pl.max(col).alias(f\"max_{col}\") for col in cols]\n        expr_last = [pl.last(col).alias(f\"last_{col}\") for col in cols]\n        #expr_first = [pl.first(col).alias(f\"first_{col}\") for col in cols]\n        expr_mean = [pl.mean(col).alias(f\"mean_{col}\") for col in cols]\n        return expr_max +expr_last+expr_mean\n    \n    def date_expr(df): \n        # 根据日期生成额外特征\n        cols = [col for col in df.columns if col[-1] in (\"D\")]\n        expr_max = [pl.max(col).alias(f\"max_{col}\") for col in cols]\n        #expr_min = [pl.min(col).alias(f\"min_{col}\") for col in cols]\n        expr_last = [pl.last(col).alias(f\"last_{col}\") for col in cols]\n        #expr_first = [pl.first(col).alias(f\"first_{col}\") for col in cols]\n        expr_mean = [pl.mean(col).alias(f\"mean_{col}\") for col in cols]\n        return  expr_max +expr_last+expr_mean\n    \n    def str_expr(df):\n        cols = [col for col in df.columns if col[-1] in (\"M\",)]\n        expr_max = [pl.max(col).alias(f\"max_{col}\") for col in cols]\n        #expr_min = [pl.min(col).alias(f\"min_{col}\") for col in cols]\n        expr_last = [pl.last(col).alias(f\"last_{col}\") for col in cols]\n        #expr_first = [pl.first(col).alias(f\"first_{col}\") for col in cols]\n        #expr_count = [pl.count(col).alias(f\"count_{col}\") for col in cols]\n        return  expr_max +expr_last#+expr_count\n    \n    def other_expr(df):\n        cols = [col for col in df.columns if col[-1] in (\"T\", \"L\")]\n        expr_max = [pl.max(col).alias(f\"max_{col}\") for col in cols]\n        #expr_min = [pl.min(col).alias(f\"min_{col}\") for col in cols]\n        expr_last = [pl.last(col).alias(f\"last_{col}\") for col in cols]\n        #expr_first = [pl.first(col).alias(f\"first_{col}\") for col in cols]\n        return  expr_max +expr_last\n    \n    def count_expr(df):\n        cols = [col for col in df.columns if \"num_group\" in col]\n        expr_max = [pl.max(col).alias(f\"max_{col}\") for col in cols] \n        #expr_min = [pl.min(col).alias(f\"min_{col}\") for col in cols]\n        expr_last = [pl.last(col).alias(f\"last_{col}\") for col in cols]\n        #expr_first = [pl.first(col).alias(f\"first_{col}\") for col in cols]\n        return  expr_max +expr_last\n    \n    def get_exprs(df):\n        exprs = Aggregator.num_expr(df) + \\\n                Aggregator.date_expr(df) + \\\n                Aggregator.str_expr(df) + \\\n                Aggregator.other_expr(df) + \\\n                Aggregator.count_expr(df)\n\n        return exprs\n\ndef read_file(path, depth=None):\n    df = pl.read_parquet(path)\n    df = df.pipe(Pipeline.set_table_dtypes)\n    if depth in [1,2]:\n        df = df.group_by(\"case_id\").agg(Aggregator.get_exprs(df)) \n    return df\n\ndef read_files(regex_path, depth=None):\n    chunks = []\n    \n    for path in glob(str(regex_path)):\n        df = pl.read_parquet(path)\n        df = df.pipe(Pipeline.set_table_dtypes)\n        if depth in [1, 2]:\n            df = df.group_by(\"case_id\").agg(Aggregator.get_exprs(df))\n        chunks.append(df)\n    \n    df = pl.concat(chunks, how=\"vertical_relaxed\")\n    df = df.unique(subset=[\"case_id\"])\n    return df\n\ndef feature_eng(df_base, depth_0, depth_1, depth_2):\n    df_base = (\n        df_base\n        .with_columns(\n            month_decision = pl.col(\"date_decision\").dt.month(),\n            weekday_decision = pl.col(\"date_decision\").dt.weekday(),\n        )\n    )\n    for i, df in enumerate(depth_0 + depth_1 + depth_2):\n        df_base = df_base.join(df, how=\"left\", on=\"case_id\", suffix=f\"_{i}\")\n    df_base = df_base.pipe(Pipeline.handle_dates)\n    return df_base\n\ndef to_pandas(df_data, cat_cols=None):\n    df_data = df_data.to_pandas()\n    if cat_cols is None:\n        cat_cols = list(df_data.select_dtypes(\"object\").columns)\n    df_data[cat_cols] = df_data[cat_cols].astype(\"category\")\n    return df_data, cat_cols\n\ndef reduce_mem_usage(df):\n    \"\"\" iterate through all the columns of a dataframe and modify the data type\n        to reduce memory usage.        \n    \"\"\"\n    start_mem = df.memory_usage().sum() / 1024**2\n    print('Memory usage of dataframe is {:.2f} MB'.format(start_mem))\n    \n    for col in df.columns:\n        col_type = df[col].dtype\n        if str(col_type)==\"category\":\n            continue\n        \n        if col_type != object:\n            c_min = df[col].min()\n            c_max = df[col].max()\n            if str(col_type)[:3] == 'int':\n                if c_min > np.iinfo(np.int8).min and c_max < np.iinfo(np.int8).max:\n                    df[col] = df[col].astype(np.int8)\n                elif c_min > np.iinfo(np.int16).min and c_max < np.iinfo(np.int16).max:\n                    df[col] = df[col].astype(np.int16)\n                elif c_min > np.iinfo(np.int32).min and c_max < np.iinfo(np.int32).max:\n                    df[col] = df[col].astype(np.int32)\n                elif c_min > np.iinfo(np.int64).min and c_max < np.iinfo(np.int64).max:\n                    df[col] = df[col].astype(np.int64)  \n            else:\n                if c_min > np.finfo(np.float16).min and c_max < np.finfo(np.float16).max:\n                    df[col] = df[col].astype(np.float16)\n                elif c_min > np.finfo(np.float32).min and c_max < np.finfo(np.float32).max:\n                    df[col] = df[col].astype(np.float32)\n                else:\n                    df[col] = df[col].astype(np.float64)\n        else:\n            continue\n    end_mem = df.memory_usage().sum() / 1024**2\n    print('Memory usage after optimization is: {:.2f} MB'.format(end_mem))\n    print('Decreased by {:.1f}%'.format(100 * (start_mem - end_mem) / start_mem))\n    \n    return df","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:35:44.872089Z","iopub.execute_input":"2024-06-06T05:35:44.872664Z","iopub.status.idle":"2024-06-06T05:35:44.911991Z","shell.execute_reply.started":"2024-06-06T05:35:44.872636Z","shell.execute_reply":"2024-06-06T05:35:44.910873Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"ROOT            = Path(\"/kaggle/input/home-credit-credit-risk-model-stability\")\n\nTRAIN_DIR       = ROOT / \"parquet_files\" / \"train\"\nTEST_DIR        = ROOT / \"parquet_files\" / \"test\"","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:35:44.913212Z","iopub.execute_input":"2024-06-06T05:35:44.913577Z","iopub.status.idle":"2024-06-06T05:35:44.924303Z","shell.execute_reply.started":"2024-06-06T05:35:44.913540Z","shell.execute_reply":"2024-06-06T05:35:44.923457Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"data_store = {\n    \"df_base\": read_file(TRAIN_DIR / \"train_base.parquet\"),\n    \"depth_0\": [\n        read_file(TRAIN_DIR / \"train_static_cb_0.parquet\"),\n        read_files(TRAIN_DIR / \"train_static_0_*.parquet\"),\n    ],\n    \"depth_1\": [\n        read_files(TRAIN_DIR / \"train_applprev_1_*.parquet\", 1),\n        read_file(TRAIN_DIR / \"train_tax_registry_a_1.parquet\", 1),\n        read_file(TRAIN_DIR / \"train_tax_registry_b_1.parquet\", 1),\n        read_file(TRAIN_DIR / \"train_tax_registry_c_1.parquet\", 1),\n        read_files(TRAIN_DIR / \"train_credit_bureau_a_1_*.parquet\", 1),\n        read_file(TRAIN_DIR / \"train_credit_bureau_b_1.parquet\", 1),\n        read_file(TRAIN_DIR / \"train_other_1.parquet\", 1),\n        read_file(TRAIN_DIR / \"train_person_1.parquet\", 1),\n        read_file(TRAIN_DIR / \"train_deposit_1.parquet\", 1),\n        read_file(TRAIN_DIR / \"train_debitcard_1.parquet\", 1),\n    ],\n    \"depth_2\": [\n        read_file(TRAIN_DIR / \"train_credit_bureau_b_2.parquet\", 2),\n        read_files(TRAIN_DIR / \"train_credit_bureau_a_2_*.parquet\", 2),\n        read_file(TRAIN_DIR / \"train_applprev_2.parquet\", 2),\n        read_file(TRAIN_DIR / \"train_person_2.parquet\", 2)\n    ]\n}","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:35:44.925473Z","iopub.execute_input":"2024-06-06T05:35:44.926222Z","iopub.status.idle":"2024-06-06T05:37:57.043919Z","shell.execute_reply.started":"2024-06-06T05:35:44.926196Z","shell.execute_reply":"2024-06-06T05:37:57.042785Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df_train = feature_eng(**data_store)\nprint(\"train data shape:\\t\", df_train.shape)\ndel data_store\ngc.collect()\ndf_train = df_train.pipe(Pipeline.filter_cols)\ndf_train, cat_cols = to_pandas(df_train)\ndf_train = reduce_mem_usage(df_train)\nprint(\"train data shape:\\t\", df_train.shape)\nnums=df_train.select_dtypes(exclude='category').columns\nfrom itertools import combinations, permutations\n#df_train=df_train[nums]\nnans_df = df_train[nums].isna()\nnans_groups={}\nfor col in nums:\n    cur_group = nans_df[col].sum()\n    try:\n        nans_groups[cur_group].append(col)\n    except:\n        nans_groups[cur_group]=[col]\ndel nans_df; x=gc.collect()\n\ndef reduce_group(grps):\n    use = []\n    for g in grps:\n        mx = 0; vx = g[0]\n        for gg in g:\n            n = df_train[gg].nunique()\n            if n>mx:\n                mx = n\n                vx = gg\n            #print(str(gg)+'-'+str(n),', ',end='')\n        use.append(vx)\n        #print()\n    print('Use these',use)\n    return use\n\ndef group_columns_by_correlation(matrix, threshold=0.8):\n    # 计算列之间的相关性\n    correlation_matrix = matrix.corr()\n\n    # 分组列\n    groups = []\n    remaining_cols = list(matrix.columns)\n    while remaining_cols:\n        col = remaining_cols.pop(0)\n        group = [col]\n        correlated_cols = [col]\n        for c in remaining_cols:\n            if correlation_matrix.loc[col, c] >= threshold:\n                group.append(c)\n                correlated_cols.append(c)\n        groups.append(group)\n        remaining_cols = [c for c in remaining_cols if c not in correlated_cols]\n    \n    return groups\n\nuses=[]\nfor k,v in nans_groups.items():\n    if len(v)>1:\n            Vs = nans_groups[k]\n            #cross_features=list(combinations(Vs, 2))\n            #make_corr(Vs)\n            grps= group_columns_by_correlation(df_train[Vs], threshold=0.8)\n            use=reduce_group(grps)\n            uses=uses+use\n            #make_corr(use)\n    else:\n        uses=uses+v\n    print('####### NAN count =',k)\nprint(uses)\nprint(len(uses))\nuses=uses+list(df_train.select_dtypes(include='category').columns)\nprint(len(uses))\ndf_train=df_train[uses]","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:37:57.046430Z","iopub.execute_input":"2024-06-06T05:37:57.047110Z","iopub.status.idle":"2024-06-06T05:39:35.774149Z","shell.execute_reply.started":"2024-06-06T05:37:57.047082Z","shell.execute_reply":"2024-06-06T05:39:35.773218Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"sample = pd.read_csv(\"/kaggle/input/home-credit-credit-risk-model-stability/sample_submission.csv\")\ndevice='gpu'\n#n_samples=200000\nn_est=6000\nDRY_RUN = True if sample.shape[0] == 10 else False   \nif DRY_RUN:\n    device='cpu'\n    df_train = df_train.iloc[:50000]\n    #n_samples=10000\n    n_est=600\nprint(device)","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:39:35.775251Z","iopub.execute_input":"2024-06-06T05:39:35.775546Z","iopub.status.idle":"2024-06-06T05:39:35.785781Z","shell.execute_reply.started":"2024-06-06T05:39:35.775518Z","shell.execute_reply":"2024-06-06T05:39:35.784864Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"data_store = {\n    \"df_base\": read_file(TEST_DIR / \"test_base.parquet\"),\n    \"depth_0\": [\n        read_file(TEST_DIR / \"test_static_cb_0.parquet\"),\n        read_files(TEST_DIR / \"test_static_0_*.parquet\"),\n    ],\n    \"depth_1\": [\n        read_files(TEST_DIR / \"test_applprev_1_*.parquet\", 1),\n        read_file(TEST_DIR / \"test_tax_registry_a_1.parquet\", 1),\n        read_file(TEST_DIR / \"test_tax_registry_b_1.parquet\", 1),\n        read_file(TEST_DIR / \"test_tax_registry_c_1.parquet\", 1),\n        read_files(TEST_DIR / \"test_credit_bureau_a_1_*.parquet\", 1),\n        read_file(TEST_DIR / \"test_credit_bureau_b_1.parquet\", 1),\n        read_file(TEST_DIR / \"test_other_1.parquet\", 1),\n        read_file(TEST_DIR / \"test_person_1.parquet\", 1),\n        read_file(TEST_DIR / \"test_deposit_1.parquet\", 1),\n        read_file(TEST_DIR / \"test_debitcard_1.parquet\", 1),\n    ],\n    \"depth_2\": [\n        read_file(TEST_DIR / \"test_credit_bureau_b_2.parquet\", 2),\n        read_files(TEST_DIR / \"test_credit_bureau_a_2_*.parquet\", 2),\n        read_file(TEST_DIR / \"test_applprev_2.parquet\", 2),\n        read_file(TEST_DIR / \"test_person_2.parquet\", 2)\n    ]\n}","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:39:35.786805Z","iopub.execute_input":"2024-06-06T05:39:35.787129Z","iopub.status.idle":"2024-06-06T05:39:36.037471Z","shell.execute_reply.started":"2024-06-06T05:39:35.787098Z","shell.execute_reply":"2024-06-06T05:39:36.036748Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df_test = feature_eng(**data_store)\nprint(\"test data shape:\\t\", df_test.shape)\ndel data_store\ngc.collect()\ndf_test = df_test.select([col for col in df_train.columns if col != \"target\"])\nprint(\"train data shape:\\t\", df_train.shape)\nprint(\"test data shape:\\t\", df_test.shape)\n\ndf_test, cat_cols = to_pandas(df_test, cat_cols)\ndf_test = reduce_mem_usage(df_test)\n\ngc.collect()\ncat_cols","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:39:36.038409Z","iopub.execute_input":"2024-06-06T05:39:36.038646Z","iopub.status.idle":"2024-06-06T05:39:36.541553Z","shell.execute_reply.started":"2024-06-06T05:39:36.038624Z","shell.execute_reply":"2024-06-06T05:39:36.540601Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### Feature Selection","metadata":{}},{"cell_type":"code","source":"y = df_train[\"target\"]\nweeks = df_train[\"WEEK_NUM\"]\ndf_train= df_train.drop(columns=[\"target\", \"case_id\", \"WEEK_NUM\"])\n\ncv = StratifiedGroupKFold(n_splits=2, shuffle=False)\n","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:39:36.542692Z","iopub.execute_input":"2024-06-06T05:39:36.542975Z","iopub.status.idle":"2024-06-06T05:39:36.672241Z","shell.execute_reply.started":"2024-06-06T05:39:36.542952Z","shell.execute_reply":"2024-06-06T05:39:36.671432Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"y","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:39:36.673305Z","iopub.execute_input":"2024-06-06T05:39:36.673624Z","iopub.status.idle":"2024-06-06T05:39:36.681126Z","shell.execute_reply.started":"2024-06-06T05:39:36.673598Z","shell.execute_reply":"2024-06-06T05:39:36.680076Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"cat_cols = [\"max_sex_738L\", \"last_relationshiptoclient_642T\",\"max_relationshiptoclient_642T\",\"max_relationshiptoclient_415T\",\"max_familystate_447L\",\"twobodfilling_608L\",\"max_incometype_1044T\",\"maritalst_385M\",\"max_empl_industry_691L\",\"max_empl_employedtotal_800L\",\"last_postype_4733339M\",\"last_rejectreason_755M\",\"max_rejectreason_755M\",\"education_1103M\",\"requesttype_4525192L\",\"credtype_322L\",\"lastrejectreason_759M\",\"last_education_1138M\",\"last_conts_type_509L\",\"inittransactioncode_186L\",\"description_5085714M\",\"max_familystate_726L\",\"max_empls_economicalst_849M\",\"last_familystate_726L\",\"max_description_351M\",\"max_education_1138M\",\"disbursementtype_67L\"]\ndf_train[cat_cols] = df_train[cat_cols].astype(str)\ndf_test[cat_cols] = df_test[cat_cols].astype(str)","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:46:26.888917Z","iopub.execute_input":"2024-06-06T05:46:26.889689Z","iopub.status.idle":"2024-06-06T05:46:26.960787Z","shell.execute_reply.started":"2024-06-06T05:46:26.889656Z","shell.execute_reply":"2024-06-06T05:46:26.960015Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df_train[\"max_sex_738L\"]","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:46:29.955903Z","iopub.execute_input":"2024-06-06T05:46:29.956562Z","iopub.status.idle":"2024-06-06T05:46:29.969503Z","shell.execute_reply.started":"2024-06-06T05:46:29.956516Z","shell.execute_reply":"2024-06-06T05:46:29.968531Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# pip install git+https://github.com/MaxHalford/xam --upgrade","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:39:36.738777Z","iopub.execute_input":"2024-06-06T05:39:36.739031Z","iopub.status.idle":"2024-06-06T05:39:36.742803Z","shell.execute_reply.started":"2024-06-06T05:39:36.739010Z","shell.execute_reply":"2024-06-06T05:39:36.741956Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import xam\nencoder = xam.feature_extraction.BayesianTargetEncoder(columns=cat_cols, prior_weight=3, suffix='')\n\n# 拟合编码器并将其应用于训练数据和测试数据\nX_train_encoded = encoder.fit_transform(df_train,y)\nX_test_encoded = encoder.transform(df_test)","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:46:34.276298Z","iopub.execute_input":"2024-06-06T05:46:34.277265Z","iopub.status.idle":"2024-06-06T05:46:34.722759Z","shell.execute_reply.started":"2024-06-06T05:46:34.277228Z","shell.execute_reply":"2024-06-06T05:46:34.721668Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"X_train_encoded[\"max_sex_738L\"]","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:48:45.785090Z","iopub.execute_input":"2024-06-06T05:48:45.785461Z","iopub.status.idle":"2024-06-06T05:48:45.793989Z","shell.execute_reply.started":"2024-06-06T05:48:45.785428Z","shell.execute_reply":"2024-06-06T05:48:45.792882Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# df_train[cat_cols] = df_train[cat_cols].astype(str)\n# df_test[cat_cols] = df_test[cat_cols].astype(str)\n# df_train = pd.concat([df_train, X_train_encoded], axis=1)\n# df_test = pd.concat([df_test, X_test_encoded], axis=1)","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:39:38.865854Z","iopub.status.idle":"2024-06-06T05:39:38.866173Z","shell.execute_reply.started":"2024-06-06T05:39:38.866019Z","shell.execute_reply":"2024-06-06T05:39:38.866031Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df_train = X_train_encoded\ndf_train[cat_cols] = df_train[cat_cols].astype(str)\ndf_test[cat_cols] = df_test[cat_cols].astype(str)","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:52:48.833926Z","iopub.execute_input":"2024-06-06T05:52:48.834757Z","iopub.status.idle":"2024-06-06T05:52:50.521273Z","shell.execute_reply.started":"2024-06-06T05:52:48.834717Z","shell.execute_reply":"2024-06-06T05:52:50.520461Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"params = {\n    \"boosting_type\": \"gbdt\",\n    \"objective\": \"binary\",\n    \"metric\": \"auc\",\n    \"max_depth\": 10,  \n    \"learning_rate\": 0.05,\n    \"n_estimators\": 2000,  \n    \"colsample_bytree\": 0.8,\n    \"colsample_bynode\": 0.8,\n    \"verbose\": -1,\n    \"random_state\": 42,\n    \"reg_alpha\": 0.1,\n    \"reg_lambda\": 10,\n    \"extra_trees\":True,\n    'num_leaves':64,\n    \"device\": device, \n    \"verbose\": -1,\n}","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:52:52.185566Z","iopub.execute_input":"2024-06-06T05:52:52.185968Z","iopub.status.idle":"2024-06-06T05:52:52.193568Z","shell.execute_reply.started":"2024-06-06T05:52:52.185936Z","shell.execute_reply":"2024-06-06T05:52:52.192788Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"n_est = 1800","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:50:09.587287Z","iopub.execute_input":"2024-06-06T05:50:09.588068Z","iopub.status.idle":"2024-06-06T05:50:09.592282Z","shell.execute_reply.started":"2024-06-06T05:50:09.588031Z","shell.execute_reply":"2024-06-06T05:50:09.591225Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from catboost import CatBoostClassifier, Pool","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:52:56.961113Z","iopub.execute_input":"2024-06-06T05:52:56.961494Z","iopub.status.idle":"2024-06-06T05:52:56.965870Z","shell.execute_reply.started":"2024-06-06T05:52:56.961455Z","shell.execute_reply":"2024-06-06T05:52:56.964883Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"%%time\n\nfitted_models_cat = []\nfitted_models_lgb = []\n\ncv_scores_cat = []\ncv_scores_lgb = []\n\nstability_results_cat = []\nstability_results_lgb = []\n\nfold = 0\nfor idx_train, idx_valid in cv.split(df_train, y, groups=weeks):#\n    \n    df_res_cat = pd.DataFrame()\n    df_res_lgb = pd.DataFrame()\n    \n    X_train, y_train = df_train.iloc[idx_train], y.iloc[idx_train]# \n    X_valid, y_valid, week_valid = df_train.iloc[idx_valid], y.iloc[idx_valid], weeks[idx_valid]\n    \n    df_res_cat['WEEK_NUM'] = list(week_valid)\n    df_res_cat['target'] = list(y_valid)\n    df_res_cat['fold'] = fold\n    \n    \n    df_res_lgb['WEEK_NUM'] = list(week_valid)\n    df_res_lgb['target'] = list(y_valid)\n    \n    fold += 1\n    \n    \n    train_pool = Pool(X_train, y_train,cat_features=cat_cols)\n    val_pool = Pool(X_valid, y_valid,cat_features=cat_cols)\n    \n    clf = CatBoostClassifier(eval_metric='AUC', task_type='GPU', learning_rate=0.03, iterations=n_est)\n    \n    random_seed=3107\n    \n    clf.fit(train_pool, eval_set=val_pool,verbose=300)\n    fitted_models_cat.append(clf)\n    y_pred_valid = clf.predict_proba(X_valid)[:,1]\n    auc_score = roc_auc_score(y_valid, y_pred_valid)\n    cv_scores_cat.append(auc_score)\n    \n    df_res_cat['score'] = list(y_pred_valid)\n    \n    \n    \n    X_train[cat_cols] = X_train[cat_cols].astype(\"category\")\n    X_valid[cat_cols] = X_valid[cat_cols].astype(\"category\")\n    \n    model = lgb.LGBMClassifier(**params)\n    model.fit(\n        X_train, y_train,\n        eval_set = [(X_valid, y_valid)],\n        callbacks = [lgb.log_evaluation(200), lgb.early_stopping(100)] )\n    \n    fitted_models_lgb.append(model)\n    y_pred_valid = model.predict_proba(X_valid)[:,1]\n    auc_score = roc_auc_score(y_valid, y_pred_valid)\n    cv_scores_lgb.append(auc_score)\n    \n    df_res_lgb['score'] = list(y_pred_valid)\n    \n    stability_results_cat.append(df_res_cat)\n    stability_results_lgb.append(df_res_lgb)\n\n\n    \n    \nprint(\"Catboost CV AUC scores: \", cv_scores_cat)\nprint(\"Maximum Catboost CV AUC score: \", max(cv_scores_cat))\n\n\nprint(\"Lightgbm CV AUC scores: \", cv_scores_lgb)\nprint(\"Maximum Lightgbm CV AUC score: \", max(cv_scores_lgb))","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:52:58.263657Z","iopub.execute_input":"2024-06-06T05:52:58.264327Z","iopub.status.idle":"2024-06-06T05:53:01.094236Z","shell.execute_reply.started":"2024-06-06T05:52:58.264291Z","shell.execute_reply":"2024-06-06T05:53:01.093299Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"df_stability_results_cat = pd.concat(stability_results_cat)\ndf_stability_results_lgb = pd.concat(stability_results_lgb)","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:39:38.876889Z","iopub.status.idle":"2024-06-06T05:39:38.877217Z","shell.execute_reply.started":"2024-06-06T05:39:38.877057Z","shell.execute_reply":"2024-06-06T05:39:38.877070Z"},"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","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:39:38.878284Z","iopub.status.idle":"2024-06-06T05:39:38.878588Z","shell.execute_reply.started":"2024-06-06T05:39:38.878436Z","shell.execute_reply":"2024-06-06T05:39:38.878448Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"try:\n    df_stability_results_cat.groupby('fold').apply(gini_stability, include_groups=False)\nexcept:\n    pass","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:39:38.879703Z","iopub.status.idle":"2024-06-06T05:39:38.880063Z","shell.execute_reply.started":"2024-06-06T05:39:38.879876Z","shell.execute_reply":"2024-06-06T05:39:38.879890Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"gini_stability(df_stability_results_cat)","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:39:38.881097Z","iopub.status.idle":"2024-06-06T05:39:38.881432Z","shell.execute_reply.started":"2024-06-06T05:39:38.881263Z","shell.execute_reply":"2024-06-06T05:39:38.881277Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"try:\n    df_stability_results_lgb.groupby('fold').apply(gini_stability, include_groups=False)\nexcept:\n    pass","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:39:38.882773Z","iopub.status.idle":"2024-06-06T05:39:38.883115Z","shell.execute_reply.started":"2024-06-06T05:39:38.882960Z","shell.execute_reply":"2024-06-06T05:39:38.882973Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"gini_stability(df_stability_results_lgb)","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:39:38.884282Z","iopub.status.idle":"2024-06-06T05:39:38.884609Z","shell.execute_reply.started":"2024-06-06T05:39:38.884447Z","shell.execute_reply":"2024-06-06T05:39:38.884461Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"class VotingModel(BaseEstimator, RegressorMixin):\n    def __init__(self, estimators):\n        super().__init__()\n        self.estimators = estimators\n        \n    def fit(self, X, y=None):\n        return self\n    \n    def predict(self, X):\n        y_preds = [estimator.predict(X) for estimator in self.estimators]\n        return np.mean(y_preds, axis=0)\n    \n    def predict_proba(self, X):\n        \n        y_preds = [estimator.predict_proba(X) for estimator in self.estimators[:5]]\n        \n        X[cat_cols] = X[cat_cols].astype(\"category\")\n        y_preds += [estimator.predict_proba(X) for estimator in self.estimators[5:]]\n        \n        return np.mean(y_preds, axis=0)\n\nmodel = VotingModel(fitted_models_cat+fitted_models_lgb)","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:39:38.885765Z","iopub.status.idle":"2024-06-06T05:39:38.886125Z","shell.execute_reply.started":"2024-06-06T05:39:38.885964Z","shell.execute_reply":"2024-06-06T05:39:38.885978Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import pandas as pd\n\n# 假设 fitted_models_cat 和 fitted_models_lgb 包含了您的 CatBoost 和 LightGBM 模型\n\n# 初始化一个空的 DataFrame 来存储特征重要性\nfeature_importances = pd.DataFrame(columns=['model', 'feature', 'importance'])\n\n# 遍历 CatBoost 模型\nfor idx, model in enumerate(fitted_models_cat):\n    importances = model.feature_importances_\n    feature_names = X_train.columns  # 假设 X_train 包含所有特征\n    df_importances = pd.DataFrame({'model': ['CatBoost'] * len(importances),\n                                   'feature': feature_names,\n                                   'importance': importances})\n    # 使用 concat 代替 append\n    feature_importances = pd.concat([feature_importances, df_importances], axis=0)\n# 遍历 LightGBM 模型\n\nfor idx, model in enumerate(fitted_models_lgb):\n    importances = model.feature_importances_\n    df_importances = pd.DataFrame({'model': ['LightGBM'] * len(importances),\n                                   'feature': feature_names,\n                                   'importance': importances})\n    # 使用 concat 代替 append\n    feature_importances = pd.concat([feature_importances, df_importances], axis=0)\n\n# 选择前 100 个重要特征\ntop_20_features = feature_importances.nlargest(100, 'importance')\n\n# 可视化特征重要性（这里只是一个示例，您可能需要根据实际情况调整）\nplt.figure(figsize=(20, 6))\nsns.barplot(x='importance', y='feature', hue='model', data=top_20_features)\nplt.title('Top 30 Feature Importances across Models')\nplt.xlabel('Importance')\nplt.ylabel('Feature')\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:39:38.887445Z","iopub.status.idle":"2024-06-06T05:39:38.887773Z","shell.execute_reply.started":"2024-06-06T05:39:38.887609Z","shell.execute_reply":"2024-06-06T05:39:38.887622Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import pandas as pd\n\n# Assuming fitted_models_cat and fitted_models_lgb contain your CatBoost and LightGBM models\n\n# Initialize an empty DataFrame to store feature importances\nfeature_importances = pd.DataFrame(columns=['model', 'feature', 'importance', 'dtype'])\n\n# Iterate through CatBoost models\nfor idx, model in enumerate(fitted_models_cat):\n    importances = model.feature_importances_\n    feature_names = X_train.columns  # Assuming X_train contains all features\n    df_importances = pd.DataFrame({'model': ['CatBoost'] * len(importances),\n                                  'feature': feature_names,\n                                  'importance': importances,\n                                  'dtype': X_train[feature_names].dtypes})\n    df_importances = df_importances[df_importances['dtype'] == 'category']\n    # Use concat instead of append\n    feature_importances = pd.concat([feature_importances, df_importances], axis=0)\n\n# Iterate through LightGBM models\nfor idx, model in enumerate(fitted_models_lgb):\n    importances = model.feature_importances_\n    df_importances = pd.DataFrame({'model': ['LightGBM'] * len(importances),\n                                  'feature': feature_names,\n                                  'importance': importances,\n                                  'dtype': X_train[feature_names].dtypes})\n    # Use concat instead of append\n    df_importances = df_importances[df_importances['dtype'] == 'category']\n    feature_importances = pd.concat([feature_importances, df_importances], axis=0)\n\n# Select top 100 features\ntop_30_features = feature_importances.nlargest(100, 'importance')\n\n# Print data types of top 30 features for each model\nprint(\"**CatBoost Model**\")\nprint(top_30_features[top_30_features['model'] == 'CatBoost'][['feature', 'dtype']].to_string(index=False))\n\nprint(\"\\n**LightGBM Model**\")\nprint(top_30_features[top_30_features['model'] == 'LightGBM'][['feature', 'dtype']].to_string(index=False))","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:39:38.889204Z","iopub.status.idle":"2024-06-06T05:39:38.889535Z","shell.execute_reply.started":"2024-06-06T05:39:38.889365Z","shell.execute_reply":"2024-06-06T05:39:38.889378Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import pandas as pd\n\n# Assuming fitted_models_cat and fitted_models_lgb contain your CatBoost and LightGBM models\n\n# Initialize an empty DataFrame to store feature importances\nfeature_importances = pd.DataFrame(columns=['model', 'feature', 'importance', 'dtype'])\n\n# Iterate through CatBoost models\nfor idx, model in enumerate(fitted_models_cat):\n    importances = model.feature_importances_\n    feature_names = X_train.columns  # Assuming X_train contains all features\n    df_importances = pd.DataFrame({'model': ['CatBoost'] * len(importances),\n                                  'feature': feature_names,\n                                  'importance': importances,\n                                  'dtype': X_train[feature_names].dtypes})\n    # Use concat instead of append\n    feature_importances = pd.concat([feature_importances, df_importances], axis=0)\n\n# Iterate through LightGBM models\nfor idx, model in enumerate(fitted_models_lgb):\n    importances = model.feature_importances_\n    df_importances = pd.DataFrame({'model': ['LightGBM'] * len(importances),\n                                  'feature': feature_names,\n                                  'importance': importances,\n                                  'dtype': X_train[feature_names].dtypes})\n    # Use concat instead of append\n    feature_importances = pd.concat([feature_importances, df_importances], axis=0)\n\n# Select features with dtype 'category'\ncat_features = feature_importances[feature_importances['dtype'] == 'category']\n\n# Print features with dtype 'category' for each model\nprint(\"**CatBoost Model (Category Features)**\")\n# print(cat_features[cat_features['model'] == 'CatBoost'][['feature', 'importance']].to_string(index=False))\nprint(cat_features[cat_features['model'] == 'CatBoost'])\nprint(\"\\n**LightGBM Model (Category Features)**\")\n# print(cat_features[cat_features['model'] == 'LightGBM'][['feature', 'importance']].to_string(index=False))","metadata":{"_kg_hide-input":true,"execution":{"iopub.status.busy":"2024-06-06T05:39:38.890670Z","iopub.status.idle":"2024-06-06T05:39:38.891017Z","shell.execute_reply.started":"2024-06-06T05:39:38.890850Z","shell.execute_reply":"2024-06-06T05:39:38.890865Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(X_train)","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:39:38.892261Z","iopub.status.idle":"2024-06-06T05:39:38.892575Z","shell.execute_reply.started":"2024-06-06T05:39:38.892421Z","shell.execute_reply":"2024-06-06T05:39:38.892434Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"print(X_train['max_relationshiptoclient_642T'])","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:39:38.893967Z","iopub.status.idle":"2024-06-06T05:39:38.894291Z","shell.execute_reply.started":"2024-06-06T05:39:38.894132Z","shell.execute_reply":"2024-06-06T05:39:38.894146Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import pandas as pd\n\n# Assuming fitted_models_cat and fitted_models_lgb contain your CatBoost and LightGBM models\n\n# Initialize an empty DataFrame to store feature importances\nfeature_importances = pd.DataFrame(columns=['model', 'feature', 'importance', 'dtype'])\n\n# Iterate through CatBoost models\nfor idx, model in enumerate(fitted_models_cat):\n  importances = model.feature_importances_\n  feature_names = X_train.columns  # Assuming X_train contains all features\n  df_importances = pd.DataFrame({'model': ['CatBoost'] * len(importances),\n                                 'feature': feature_names,\n                                 'importance': importances,\n                                 'dtype': X_train[feature_names].dtypes})\n  # Filter for category features before concat\n  df_importances = df_importances[df_importances['dtype'] == 'category']\n  # Use concat instead of append\n  feature_importances = pd.concat([feature_importances, df_importances], axis=0)\n\n# Iterate through LightGBM models\nfor idx, model in enumerate(fitted_models_lgb):\n  importances = model.feature_importances_\n  df_importances = pd.DataFrame({'model': ['LightGBM'] * len(importances),\n                                 'feature': feature_names,\n                                 'importance': importances,\n                                 'dtype': X_train[feature_names].dtypes})\n  # Filter for category features before concat\n  df_importances = df_importances[df_importances['dtype'] == 'category']\n  # Use concat instead of append\n  feature_importances = pd.concat([feature_importances, df_importances], axis=0)\n\n# No need to select top 100 features, filter within the loop\n\n# Print data types of category features for each model\nprint(\"**CatBoost Model**\")\nprint(feature_importances[feature_importances['model'] == 'CatBoost'][['feature', 'dtype']].to_string(index=False))\n\nprint(\"\\n**LightGBM Model**\")\nprint(feature_importances[feature_importances['model'] == 'LightGBM'][['feature', 'dtype']].to_string(index=False))\n","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:39:38.895721Z","iopub.status.idle":"2024-06-06T05:39:38.896058Z","shell.execute_reply.started":"2024-06-06T05:39:38.895898Z","shell.execute_reply":"2024-06-06T05:39:38.895912Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"### Submission","metadata":{}},{"cell_type":"code","source":"df_test = df_test.drop(columns=[\"WEEK_NUM\"])\ndf_test = df_test.set_index(\"case_id\")\n\n\ny_pred = pd.Series(model.predict_proba(df_test)[:, 1], index=df_test.index)\ndf_subm = pd.read_csv(ROOT / \"sample_submission.csv\")\ndf_subm = df_subm.set_index(\"case_id\")\n\ndf_subm[\"score\"] = y_pred\ndf_subm.to_csv(\"submission.csv\")\ndf_subm","metadata":{"execution":{"iopub.status.busy":"2024-06-06T05:39:38.897060Z","iopub.status.idle":"2024-06-06T05:39:38.897359Z","shell.execute_reply.started":"2024-06-06T05:39:38.897209Z","shell.execute_reply":"2024-06-06T05:39:38.897221Z"},"trusted":true},"execution_count":null,"outputs":[]}]}