{"metadata":{"kernelspec":{"language":"python","display_name":"Python 3","name":"python3"},"language_info":{"pygments_lexer":"ipython3","nbconvert_exporter":"python","version":"3.6.4","file_extension":".py","codemirror_mode":{"name":"ipython","version":3},"name":"python","mimetype":"text/x-python"}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"# XGBoost Starter - LB 0.793\nIn this notebook we build and train an XGBoost model using @raddar Kaggle dataset from [here][1] with discussion [here][2]. Then we engineer features suggested by @huseyincot in his notebooks [here][3] and [here][4]. This XGB model achieves CV 0.792 LB 0.793! When training with XGB, we use a special XGB dataloader called `DeviceQuantileDMatrix` which uses a small GPU memory footprint. This allows us to engineer more additional columns and train with more rows of data. Our feature engineering is performed using [RAPIDS][5] on the GPU to create new features quickly.\n\n[1]: https://www.kaggle.com/datasets/raddar/amex-data-integer-dtypes-parquet-format\n[2]: https://www.kaggle.com/competitions/amex-default-prediction/discussion/328514\n[3]: https://www.kaggle.com/code/huseyincot/amex-catboost-0-793\n[4]: https://www.kaggle.com/code/huseyincot/amex-agg-data-how-it-created\n[5]: https://rapids.ai/","metadata":{}},{"cell_type":"markdown","source":"# Load Libraries","metadata":{}},{"cell_type":"code","source":"# LOAD LIBRARIES\nimport pandas as pd, numpy as np # CPU libraries\nimport cupy, cudf # GPU libraries\nimport matplotlib.pyplot as plt, gc, os\n\nfrom tsfresh.feature_extraction import feature_calculators as fc\n\nprint('RAPIDS version',cudf.__version__)","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:56:28.460386Z","iopub.execute_input":"2023-05-04T13:56:28.460946Z","iopub.status.idle":"2023-05-04T13:56:34.043277Z","shell.execute_reply.started":"2023-05-04T13:56:28.460835Z","shell.execute_reply":"2023-05-04T13:56:34.041698Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# VERSION NAME FOR SAVED MODEL FILES\nVER = 1\n\n# TRAIN RANDOM SEED\nSEED = 42\n\n# FILL NAN VALUE\nNAN_VALUE = -127 # will fit in int8\n\n# FOLDS PER MODEL\nFOLDS = 5","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:56:34.047786Z","iopub.execute_input":"2023-05-04T13:56:34.048236Z","iopub.status.idle":"2023-05-04T13:56:34.056953Z","shell.execute_reply.started":"2023-05-04T13:56:34.048199Z","shell.execute_reply":"2023-05-04T13:56:34.056120Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Process and Feature Engineer Train Data\nWe will load @raddar Kaggle dataset from [here][1] with discussion [here][2]. Then we will engineer features suggested by @huseyincot in his notebooks [here][3] and [here][4]. We will use [RAPIDS][5] and the GPU to create new features quickly.\n\n[1]: https://www.kaggle.com/datasets/raddar/amex-data-integer-dtypes-parquet-format\n[2]: https://www.kaggle.com/competitions/amex-default-prediction/discussion/328514\n[3]: https://www.kaggle.com/code/huseyincot/amex-catboost-0-793\n[4]: https://www.kaggle.com/code/huseyincot/amex-agg-data-how-it-created\n[5]: https://rapids.ai/","metadata":{}},{"cell_type":"code","source":"def read_file(path = '', usecols = None):\n    # LOAD DATAFRAME\n    if usecols is not None: df = cudf.read_parquet(path, columns=usecols)\n    else: df = cudf.read_parquet(path)\n    # REDUCE DTYPE FOR CUSTOMER AND DATE\n    df['customer_ID'] = df['customer_ID'].str[-16:].str.hex_to_int().astype('int64')\n    df.S_2 = cudf.to_datetime( df.S_2 )\n    # SORT BY CUSTOMER AND DATE (so agg('last') works correctly)\n    #df = df.sort_values(['customer_ID','S_2'])\n    #df = df.reset_index(drop=True)\n    # FILL NAN\n    df = df.fillna(NAN_VALUE) \n    print('shape of data:', df.shape)\n    \n    return df\n\nprint('Reading train data...')\nTRAIN_PATH = '../input/amex-data-integer-dtypes-parquet-format/train.parquet'\ntrain = read_file(path = TRAIN_PATH)","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:56:34.061252Z","iopub.execute_input":"2023-05-04T13:56:34.061997Z","iopub.status.idle":"2023-05-04T13:57:00.063928Z","shell.execute_reply.started":"2023-05-04T13:56:34.061961Z","shell.execute_reply":"2023-05-04T13:57:00.063079Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# train = train.to_pandas()","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:57:00.066147Z","iopub.execute_input":"2023-05-04T13:57:00.066689Z","iopub.status.idle":"2023-05-04T13:57:00.070929Z","shell.execute_reply.started":"2023-05-04T13:57:00.066652Z","shell.execute_reply":"2023-05-04T13:57:00.069769Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train = train[:10000]","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:57:00.072286Z","iopub.execute_input":"2023-05-04T13:57:00.072726Z","iopub.status.idle":"2023-05-04T13:57:00.091614Z","shell.execute_reply.started":"2023-05-04T13:57:00.072692Z","shell.execute_reply":"2023-05-04T13:57:00.090958Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train.head()","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:57:00.092769Z","iopub.execute_input":"2023-05-04T13:57:00.093133Z","iopub.status.idle":"2023-05-04T13:57:00.306476Z","shell.execute_reply.started":"2023-05-04T13:57:00.093100Z","shell.execute_reply":"2023-05-04T13:57:00.305621Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"all_cols0 = [c for c in list(train.columns)]\nall_cols = [c for c in all_cols0 if c not in ['customer_ID','S_2']]\ncat_features = [\"B_30\",\"B_38\",\"D_114\",\"D_116\",\"D_117\",\"D_120\",\"D_126\",\"D_63\",\"D_64\",\"D_66\",\"D_68\"]\nnum_features = [col for col in all_cols if col not in cat_features]\nid_num_features = [col for col in all_cols0 if col not in cat_features]\n\ntsfresh_train = train[id_num_features]\ntsfresh_train = tsfresh_train.to_pandas()","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:57:00.307993Z","iopub.execute_input":"2023-05-04T13:57:00.308408Z","iopub.status.idle":"2023-05-04T13:57:00.389677Z","shell.execute_reply.started":"2023-05-04T13:57:00.308367Z","shell.execute_reply":"2023-05-04T13:57:00.388887Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"from tsfresh import extract_features\n\nfc_parameters = {\n    \"abs_energy\": None,\n    \"count_above_mean\": None,\n    \"count_below_mean\": None,\n    \"mean_abs_change\": None,\n    \"mean_change\": None\n}\n\nX = extract_features(tsfresh_train, column_id='customer_ID', column_sort='S_2', default_fc_parameters = fc_parameters)\nX.index.name = 'customer_ID'","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:57:00.391152Z","iopub.execute_input":"2023-05-04T13:57:00.391505Z","iopub.status.idle":"2023-05-04T13:57:33.565843Z","shell.execute_reply.started":"2023-05-04T13:57:00.391471Z","shell.execute_reply":"2023-05-04T13:57:33.565016Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"X","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:57:33.567312Z","iopub.execute_input":"2023-05-04T13:57:33.567734Z","iopub.status.idle":"2023-05-04T13:57:33.604669Z","shell.execute_reply.started":"2023-05-04T13:57:33.567681Z","shell.execute_reply":"2023-05-04T13:57:33.603843Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"X = cudf.from_pandas(X)","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:57:33.608081Z","iopub.execute_input":"2023-05-04T13:57:33.608466Z","iopub.status.idle":"2023-05-04T13:57:34.050851Z","shell.execute_reply.started":"2023-05-04T13:57:33.608427Z","shell.execute_reply":"2023-05-04T13:57:34.049960Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"def process_and_feature_engineer(df):\n    # FEATURE ENGINEERING FROM \n    # https://www.kaggle.com/code/huseyincot/amex-agg-data-how-it-created\n    all_cols = [c for c in list(df.columns) if c not in ['customer_ID','S_2']]\n    cat_features = [\"B_30\",\"B_38\",\"D_114\",\"D_116\",\"D_117\",\"D_120\",\"D_126\",\"D_63\",\"D_64\",\"D_66\",\"D_68\"]\n    num_features = [col for col in all_cols if col not in cat_features]\n\n    test_num_agg = df.groupby(\"customer_ID\")[num_features].agg(['mean', 'std', 'min', 'max', 'last'])\n    test_num_agg.columns = ['_'.join(x) for x in test_num_agg.columns]\n\n    test_cat_agg = df.groupby(\"customer_ID\")[cat_features].agg(['count', 'last', 'nunique'])\n    test_cat_agg.columns = ['_'.join(x) for x in test_cat_agg.columns]\n\n    df = cudf.concat([test_num_agg, test_cat_agg], axis=1)\n    del test_num_agg, test_cat_agg\n    print('shape after engineering', df.shape )\n    \n    return df\n\ntrain = process_and_feature_engineer(train)","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:57:34.052095Z","iopub.execute_input":"2023-05-04T13:57:34.052461Z","iopub.status.idle":"2023-05-04T13:57:34.243308Z","shell.execute_reply.started":"2023-05-04T13:57:34.052425Z","shell.execute_reply":"2023-05-04T13:57:34.242280Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"train.head()","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:57:34.244878Z","iopub.execute_input":"2023-05-04T13:57:34.245292Z","iopub.status.idle":"2023-05-04T13:57:35.089484Z","shell.execute_reply.started":"2023-05-04T13:57:34.245252Z","shell.execute_reply":"2023-05-04T13:57:35.088658Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# ADD TARGETS\ntargets = cudf.read_csv('../input/amex-default-prediction/train_labels.csv')\ntargets['customer_ID'] = targets['customer_ID'].str[-16:].str.hex_to_int().astype('int64')\ntargets = targets.set_index('customer_ID')\ntrain = train.merge(X, left_index=True, right_index=True, how='left')\ntrain = train.merge(targets, left_index=True, right_index=True, how='left')\ntrain.target = train.target.astype('int8')\ndel X, targets\n\n# NEEDED TO MAKE CV DETERMINISTIC (cudf merge above randomly shuffles rows)\ntrain = train.sort_index().reset_index()\n\n# FEATURES\nFEATURES = train.columns[1:-1]\nprint(f'There are {len(FEATURES)} features!')","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:57:35.090885Z","iopub.execute_input":"2023-05-04T13:57:35.091250Z","iopub.status.idle":"2023-05-04T13:57:36.223704Z","shell.execute_reply.started":"2023-05-04T13:57:35.091215Z","shell.execute_reply":"2023-05-04T13:57:36.222830Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Train XGB\nWe will train using `DeviceQuantileDMatrix`. This has a very small GPU memory footprint.","metadata":{}},{"cell_type":"code","source":"# LOAD XGB LIBRARY\nfrom sklearn.model_selection import KFold\nimport xgboost as xgb\nprint('XGB Version',xgb.__version__)\n\n# XGB MODEL PARAMETERS\nxgb_parms = { \n    'max_depth':4, \n    'learning_rate':0.05, \n    'subsample':0.8,\n    'colsample_bytree':0.6, \n    'eval_metric':'logloss',\n    'objective':'binary:logistic',\n    'tree_method':'gpu_hist',\n    'predictor':'gpu_predictor',\n    'random_state':SEED\n}","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:57:36.225049Z","iopub.execute_input":"2023-05-04T13:57:36.225392Z","iopub.status.idle":"2023-05-04T13:57:36.311073Z","shell.execute_reply.started":"2023-05-04T13:57:36.225358Z","shell.execute_reply":"2023-05-04T13:57:36.310138Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"import lightgbm as lgb\n\nlgb_parms = { \n    'max_depth':4, \n    'learning_rate':0.05, \n    'subsample':0.8,\n    'colsample_bytree':0.6, \n    'metric':'logloss',\n    'objective':'binary',\n    'random_state':SEED,\n    'verbose':-1\n}","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:57:36.312427Z","iopub.execute_input":"2023-05-04T13:57:36.312879Z","iopub.status.idle":"2023-05-04T13:57:36.589525Z","shell.execute_reply.started":"2023-05-04T13:57:36.312838Z","shell.execute_reply":"2023-05-04T13:57:36.588721Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# NEEDED WITH DeviceQuantileDMatrix BELOW\nclass IterLoadForDMatrix(xgb.core.DataIter):\n    def __init__(self, df=None, features=None, target=None, batch_size=256*1024):\n        self.features = features\n        self.target = target\n        self.df = df\n        self.it = 0 # set iterator to 0\n        self.batch_size = batch_size\n        self.batches = int( np.ceil( len(df) / self.batch_size ) )\n        super().__init__()\n\n    def reset(self):\n        '''Reset the iterator'''\n        self.it = 0\n\n    def next(self, input_data):\n        '''Yield next batch of data.'''\n        if self.it == self.batches:\n            return 0 # Return 0 when there's no more batch.\n        \n        a = self.it * self.batch_size\n        b = min( (self.it + 1) * self.batch_size, len(self.df) )\n        dt = cudf.DataFrame(self.df.iloc[a:b])\n        input_data(data=dt[self.features], label=dt[self.target]) #, weight=dt['weight'])\n        self.it += 1\n        return 1","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:57:36.590903Z","iopub.execute_input":"2023-05-04T13:57:36.591398Z","iopub.status.idle":"2023-05-04T13:57:36.602526Z","shell.execute_reply.started":"2023-05-04T13:57:36.591356Z","shell.execute_reply":"2023-05-04T13:57:36.601670Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# https://www.kaggle.com/kyakovlev\n# https://www.kaggle.com/competitions/amex-default-prediction/discussion/327534\ndef amex_metric_mod(y_true, y_pred):\n\n    labels     = np.transpose(np.array([y_true, y_pred]))\n    labels     = labels[labels[:, 1].argsort()[::-1]]\n    weights    = np.where(labels[:,0]==0, 20, 1)\n    cut_vals   = labels[np.cumsum(weights) <= int(0.04 * np.sum(weights))]\n    top_four   = np.sum(cut_vals[:,0]) / np.sum(labels[:,0])\n\n    gini = [0,0]\n    for i in [1,0]:\n        labels         = np.transpose(np.array([y_true, y_pred]))\n        labels         = labels[labels[:, i].argsort()[::-1]]\n        weight         = np.where(labels[:,0]==0, 20, 1)\n        weight_random  = np.cumsum(weight / np.sum(weight))\n        total_pos      = np.sum(labels[:, 0] *  weight)\n        cum_pos_found  = np.cumsum(labels[:, 0] * weight)\n        lorentz        = cum_pos_found / total_pos\n        gini[i]        = np.sum((lorentz - weight_random) * weight)\n\n    return 0.5 * (gini[1]/gini[0] + top_four)","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:57:36.603454Z","iopub.execute_input":"2023-05-04T13:57:36.604462Z","iopub.status.idle":"2023-05-04T13:57:36.615192Z","shell.execute_reply.started":"2023-05-04T13:57:36.604434Z","shell.execute_reply":"2023-05-04T13:57:36.614336Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# importances = []\n# oof = []\n# train = train.to_pandas() # free GPU memory\n# TRAIN_SUBSAMPLE = 1.0\n# gc.collect()\n\n# skf = KFold(n_splits=FOLDS, shuffle=True, random_state=SEED)\n# for fold,(train_idx, valid_idx) in enumerate(skf.split(\n#             train, train.target )):\n    \n#     # TRAIN WITH SUBSAMPLE OF TRAIN FOLD DATA\n#     if TRAIN_SUBSAMPLE<1.0:\n#         np.random.seed(SEED)\n#         train_idx = np.random.choice(train_idx, \n#                        int(len(train_idx)*TRAIN_SUBSAMPLE), replace=False)\n#         np.random.seed(None)\n    \n#     print('#'*25)\n#     print('### Fold',fold+1)\n#     print('### Train size',len(train_idx),'Valid size',len(valid_idx))\n#     print(f'### Training with {int(TRAIN_SUBSAMPLE*100)}% fold data...')\n#     print('#'*25)\n    \n#     # TRAIN, VALID, TEST FOR FOLD K\n#     Xy_train = IterLoadForDMatrix(train.loc[train_idx], FEATURES, 'target')\n#     X_valid = train.loc[valid_idx, FEATURES]\n#     y_valid = train.loc[valid_idx, 'target']\n    \n#     dtrain = xgb.DeviceQuantileDMatrix(Xy_train, max_bin=256)\n#     dvalid = xgb.DMatrix(data=X_valid, label=y_valid)\n    \n#     # TRAIN MODEL FOLD K\n#     model = xgb.train(xgb_parms, \n#                 dtrain=dtrain,\n#                 evals=[(dtrain,'train'),(dvalid,'valid')],\n#                 num_boost_round=9999,\n#                 early_stopping_rounds=100,\n#                 verbose_eval=100) \n#     model.save_model(f'XGB_v{VER}_fold{fold}.xgb')\n    \n#     # GET FEATURE IMPORTANCE FOR FOLD K\n#     dd = model.get_score(importance_type='weight')\n#     df = pd.DataFrame({'feature':dd.keys(),f'importance_{fold}':dd.values()})\n#     importances.append(df)\n            \n#     # INFER OOF FOLD K\n#     oof_preds = model.predict(dvalid)\n#     acc = amex_metric_mod(y_valid.values, oof_preds)\n#     print('Kaggle Metric =',acc,'\\n')\n    \n#     # SAVE OOF\n#     df = train.loc[valid_idx, ['customer_ID','target'] ].copy()\n#     df['oof_pred'] = oof_preds\n#     oof.append( df )\n    \n#     del dtrain, Xy_train, dd, df\n#     del X_valid, y_valid, dvalid, model\n#     _ = gc.collect()\n    \n# print('#'*25)\n# oof = pd.concat(oof,axis=0,ignore_index=True).set_index('customer_ID')\n# acc = amex_metric_mod(oof.target.values, oof.oof_pred.values)\n# print('OVERALL CV Kaggle Metric =',acc)","metadata":{"scrolled":true,"execution":{"iopub.status.busy":"2023-05-04T13:57:36.616616Z","iopub.execute_input":"2023-05-04T13:57:36.617075Z","iopub.status.idle":"2023-05-04T13:57:36.628507Z","shell.execute_reply.started":"2023-05-04T13:57:36.617039Z","shell.execute_reply":"2023-05-04T13:57:36.627770Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"importances = []\noof = []\ntrain = train.to_pandas() # free GPU memory\nTRAIN_SUBSAMPLE = 1.0\ngc.collect()\n\nskf = KFold(n_splits=FOLDS, shuffle=True, random_state=SEED)\nfor fold,(train_idx, valid_idx) in enumerate(skf.split(\n            train, train.target )):\n    \n    # TRAIN WITH SUBSAMPLE OF TRAIN FOLD DATA\n    if TRAIN_SUBSAMPLE<1.0:\n        np.random.seed(SEED)\n        train_idx = np.random.choice(train_idx, \n                       int(len(train_idx)*TRAIN_SUBSAMPLE), replace=False)\n        np.random.seed(None)\n    \n    print('#'*25)\n    print('### Fold',fold+1)\n    print('### Train size',len(train_idx),'Valid size',len(valid_idx))\n    print(f'### Training with {int(TRAIN_SUBSAMPLE*100)}% fold data...')\n    print('#'*25)\n    \n    # TRAIN, VALID, TEST FOR FOLD K\n#     Xy_train = IterLoadForDMatrix(train.loc[train_idx], FEATURES, 'target')\n    X_train = train.loc[train_idx, FEATURES]\n    y_train = train.loc[train_idx, 'target']\n    X_valid = train.loc[valid_idx, FEATURES]\n    y_valid = train.loc[valid_idx, 'target']\n    \n#     dtrain = xgb.DeviceQuantileDMatrix(Xy_train, max_bin=256)\n#     dvalid = xgb.DMatrix(data=X_valid, label=y_valid)\n\n    lgtrain = lgb.Dataset(X_train, label=y_train)\n    lgval = lgb.Dataset(X_valid, label=y_valid)\n    \n    # TRAIN MODEL FOLD K\n#     model = xgb.train(xgb_parms, \n#                 dtrain=dtrain,\n#                 evals=[(dtrain,'train'),(dvalid,'valid')],\n#                 num_boost_round=9999,\n#                 early_stopping_rounds=100,\n#                 verbose_eval=100)\n\n#     model.save_model(f'XGB_v{VER}_fold{fold}.xgb')\n    \n    model = lgb.train(lgb_parms, \n            lgtrain,\n            100,\n            valid_sets=[lgval])\n    \n    model.save_model(f'LGB_v{VER}_fold{fold}.txt', num_iteration=model.best_iteration)\n    \n    # GET FEATURE IMPORTANCE FOR FOLD K\n#     dd = model.get_score(importance_type='weight')\n#     df = pd.DataFrame({'feature':dd.keys(),f'importance_{fold}':dd.values()})\n#     importances.append(df)\n            \n    # INFER OOF FOLD K\n    oof_preds = model.predict(X_valid)\n    acc = amex_metric_mod(y_valid.values, oof_preds)\n    print('Kaggle Metric =',acc,'\\n')\n    \n    # SAVE OOF\n    df = train.loc[valid_idx, ['customer_ID','target'] ].copy()\n    df['oof_pred'] = oof_preds\n    oof.append( df )\n    \n    del X_train, y_train, df\n    del X_valid, y_valid, model\n    _ = gc.collect()\n    \nprint('#'*25)\noof = pd.concat(oof,axis=0,ignore_index=True).set_index('customer_ID')\nacc = amex_metric_mod(oof.target.values, oof.oof_pred.values)\nprint('OVERALL CV Kaggle Metric =',acc)","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:57:36.629848Z","iopub.execute_input":"2023-05-04T13:57:36.630341Z","iopub.status.idle":"2023-05-04T13:57:46.005893Z","shell.execute_reply.started":"2023-05-04T13:57:36.630289Z","shell.execute_reply":"2023-05-04T13:57:46.004083Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# CLEAN RAM\ndel train\n_ = gc.collect()","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:57:46.007142Z","iopub.execute_input":"2023-05-04T13:57:46.007594Z","iopub.status.idle":"2023-05-04T13:57:46.191750Z","shell.execute_reply.started":"2023-05-04T13:57:46.007556Z","shell.execute_reply":"2023-05-04T13:57:46.190770Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Save OOF Preds","metadata":{}},{"cell_type":"code","source":"oof_xgb = pd.read_parquet(TRAIN_PATH, columns=['customer_ID']).drop_duplicates()\noof_xgb['customer_ID_hash'] = oof_xgb['customer_ID'].apply(lambda x: int(x[-16:],16) ).astype('int64')\noof_xgb = oof_xgb.set_index('customer_ID_hash')\noof_xgb = oof_xgb.merge(oof, left_index=True, right_index=True)\noof_xgb = oof_xgb.sort_index().reset_index(drop=True)\noof_xgb.to_csv(f'oof_xgb_v{VER}.csv',index=False)\noof_xgb.head()","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:57:46.193343Z","iopub.execute_input":"2023-05-04T13:57:46.193973Z","iopub.status.idle":"2023-05-04T13:57:48.675484Z","shell.execute_reply.started":"2023-05-04T13:57:46.193924Z","shell.execute_reply":"2023-05-04T13:57:48.674640Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# PLOT OOF PREDICTIONS\nplt.hist(oof_xgb.oof_pred.values, bins=100)\nplt.title('OOF Predictions')\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:57:48.676990Z","iopub.execute_input":"2023-05-04T13:57:48.677563Z","iopub.status.idle":"2023-05-04T13:57:49.035905Z","shell.execute_reply.started":"2023-05-04T13:57:48.677519Z","shell.execute_reply":"2023-05-04T13:57:49.034108Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# CLEAR VRAM, RAM FOR INFERENCE BELOW\ndel oof_xgb, oof\n_ = gc.collect()","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:57:49.037140Z","iopub.execute_input":"2023-05-04T13:57:49.037577Z","iopub.status.idle":"2023-05-04T13:57:49.219425Z","shell.execute_reply.started":"2023-05-04T13:57:49.037540Z","shell.execute_reply":"2023-05-04T13:57:49.218340Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Feature Importance","metadata":{}},{"cell_type":"code","source":"# import matplotlib.pyplot as plt\n\n# df = importances[0].copy()\n# for k in range(1,FOLDS): df = df.merge(importances[k], on='feature', how='left')\n# df['importance'] = df.iloc[:,1:].mean(axis=1)\n# df = df.sort_values('importance',ascending=False)\n# df.to_csv(f'xgb_feature_importance_v{VER}.csv',index=False)","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:57:49.221506Z","iopub.execute_input":"2023-05-04T13:57:49.221959Z","iopub.status.idle":"2023-05-04T13:57:49.229276Z","shell.execute_reply.started":"2023-05-04T13:57:49.221922Z","shell.execute_reply":"2023-05-04T13:57:49.228401Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# NUM_FEATURES = 20\n# plt.figure(figsize=(10,5*NUM_FEATURES//10))\n# plt.barh(np.arange(NUM_FEATURES,0,-1), df.importance.values[:NUM_FEATURES])\n# plt.yticks(np.arange(NUM_FEATURES,0,-1), df.feature.values[:NUM_FEATURES])\n# plt.title(f'XGB Feature Importance - Top {NUM_FEATURES}')\n# plt.show()","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:57:49.230411Z","iopub.execute_input":"2023-05-04T13:57:49.230899Z","iopub.status.idle":"2023-05-04T13:57:49.238574Z","shell.execute_reply.started":"2023-05-04T13:57:49.230865Z","shell.execute_reply":"2023-05-04T13:57:49.237721Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Process and Feature Engineer Test Data\nWe will load @raddar Kaggle dataset from [here][1] with discussion [here][2]. Then we will engineer features suggested by @huseyincot in his notebooks [here][1] and [here][4]. We will use [RAPIDS][5] and the GPU to create new features quickly.\n\n[1]: https://www.kaggle.com/datasets/raddar/amex-data-integer-dtypes-parquet-format\n[2]: https://www.kaggle.com/competitions/amex-default-prediction/discussion/328514\n[3]: https://www.kaggle.com/code/huseyincot/amex-catboost-0-793\n[4]: https://www.kaggle.com/code/huseyincot/amex-agg-data-how-it-created\n[5]: https://rapids.ai/","metadata":{}},{"cell_type":"code","source":"# CALCULATE SIZE OF EACH SEPARATE TEST PART\ndef get_rows(customers, test, NUM_PARTS = 4, verbose = ''):\n    chunk = len(customers)//NUM_PARTS\n    if verbose != '':\n        print(f'We will process {verbose} data as {NUM_PARTS} separate parts.')\n        print(f'There will be {chunk} customers in each part (except the last part).')\n        print('Below are number of rows in each part:')\n    rows = []\n\n    for k in range(NUM_PARTS):\n        if k==NUM_PARTS-1: cc = customers[k*chunk:]\n        else: cc = customers[k*chunk:(k+1)*chunk]\n        s = test.loc[test.customer_ID.isin(cc)].shape[0]\n        rows.append(s)\n    if verbose != '': print( rows )\n    return rows,chunk\n\n# COMPUTE SIZE OF 4 PARTS FOR TEST DATA\nNUM_PARTS = 4\nTEST_PATH = '../input/amex-data-integer-dtypes-parquet-format/test.parquet'\n\nprint(f'Reading test data...')\ntest = read_file(path = TEST_PATH, usecols = ['customer_ID','S_2'])\ncustomers = test[['customer_ID']].drop_duplicates().sort_index().values.flatten()\nrows,num_cust = get_rows(customers, test[['customer_ID']], NUM_PARTS = NUM_PARTS, verbose = 'test')","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:57:49.239641Z","iopub.execute_input":"2023-05-04T13:57:49.240563Z","iopub.status.idle":"2023-05-04T13:57:51.931892Z","shell.execute_reply.started":"2023-05-04T13:57:49.240529Z","shell.execute_reply":"2023-05-04T13:57:51.931049Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Infer Test","metadata":{}},{"cell_type":"code","source":"# INFER TEST DATA IN PARTS\nskip_rows = 0\nskip_cust = 0\ntest_preds = []\n\nfor k in range(NUM_PARTS):\n    \n    # READ PART OF TEST DATA\n    print(f'\\nReading test data...')\n    test = read_file(path = TEST_PATH)\n    test = test.iloc[skip_rows:skip_rows+rows[k]]\n    skip_rows += rows[k]\n    print(f'=> Test part {k+1} has shape', test.shape )\n    \n#     test = test[:100]\n    \n    tsfresh_test = test[id_num_features]\n    tsfresh_test = tsfresh_test.to_pandas()\n    X = extract_features(tsfresh_test, column_id='customer_ID', column_sort='S_2', default_fc_parameters = fc_parameters)\n    X.index.name = 'customer_ID'\n    \n    X = cudf.from_pandas(X)\n    \n    # PROCESS AND FEATURE ENGINEER PART OF TEST DATA\n    test = process_and_feature_engineer(test)\n    \n    test = test.merge(X, left_index=True, right_index=True, how='left')\n    del X\n    \n    test = test.sort_index().reset_index()\n    \n    if k==NUM_PARTS-1: test = test.loc[customers[skip_cust:]]\n    else: test = test.loc[customers[skip_cust:skip_cust+num_cust]]\n    skip_cust += num_cust\n    \n    # TEST DATA FOR XGB\n    X_test = test[FEATURES]\n    lgtest = lgb.Dataset(X_test)\n#     dtest = xgb.DMatrix(data=X_test)\n    test = test[['P_2_mean']] # reduce memory\n    del X_test\n    gc.collect()\n\n    # INFER XGB MODELS ON TEST DATA\n    model = lgb.Booster(model_file=f'LGB_v{VER}_fold0.txt')\n#     model.save_model(f'LGB_v{VER}_fold{fold}.txt', num_iteration=model.best_iteration)\n#     model.load_model(f'XGB_v{VER}_fold0.xgb')\n    preds = model.predict(lgtest)\n    for f in range(1,FOLDS):\n        model = lgb.Booster(model_file=f'LGB_v{VER}_fold{f}.txt')\n        preds += model.predict(lgtest)\n    preds /= FOLDS\n    test_preds.append(preds)\n\n    # CLEAN MEMORY\n    del dtest, model\n    _ = gc.collect()","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:57:51.933432Z","iopub.execute_input":"2023-05-04T13:57:51.933845Z","iopub.status.idle":"2023-05-04T13:58:49.319730Z","shell.execute_reply.started":"2023-05-04T13:57:51.933809Z","shell.execute_reply":"2023-05-04T13:58:49.318481Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"markdown","source":"# Create Submission CSV","metadata":{}},{"cell_type":"code","source":"# WRITE SUBMISSION FILE\ntest_preds = np.concatenate(test_preds)\ntest = cudf.DataFrame(index=customers,data={'prediction':test_preds})\nsub = cudf.read_csv('../input/amex-default-prediction/sample_submission.csv')[['customer_ID']]\nsub['customer_ID_hash'] = sub['customer_ID'].str[-16:].str.hex_to_int().astype('int64')\nsub = sub.set_index('customer_ID_hash')\nsub = sub.merge(test[['prediction']], left_index=True, right_index=True, how='left')\nsub = sub.reset_index(drop=True)\n\n# DISPLAY PREDICTIONS\nsub.to_csv(f'submission_xgb_v{VER}.csv',index=False)\nprint('Submission file shape is', sub.shape )\nsub.head()","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:58:49.320985Z","iopub.status.idle":"2023-05-04T13:58:49.321446Z","shell.execute_reply.started":"2023-05-04T13:58:49.321221Z","shell.execute_reply":"2023-05-04T13:58:49.321245Z"},"trusted":true},"execution_count":null,"outputs":[]},{"cell_type":"code","source":"# PLOT PREDICTIONS\nplt.hist(sub.to_pandas().prediction, bins=100)\nplt.title('Test Predictions')\nplt.show()","metadata":{"execution":{"iopub.status.busy":"2023-05-04T13:58:49.322686Z","iopub.status.idle":"2023-05-04T13:58:49.323310Z","shell.execute_reply.started":"2023-05-04T13:58:49.323065Z","shell.execute_reply":"2023-05-04T13:58:49.323089Z"},"trusted":true},"execution_count":null,"outputs":[]}]}