{"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":"nvidiaTeslaT4","dataSources":[{"sourceId":84896,"databundleVersionId":10305135,"sourceType":"competition"},{"sourceId":216741403,"sourceType":"kernelVersion"}],"isInternetEnabled":false,"language":"python","sourceType":"notebook","isGpuEnabled":true}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"markdown","source":"# Introduce","metadata":{}},{"cell_type":"markdown","source":"

\n The purpose of this notebook is, I want to know where is the gap between me and the first place? I conducted research by trying different ways to compare CV scores.\n

\n\n

\n If you want to boost your score even further, try improving your CV and LB by reducing your learning rate by a factor of 10 and increasing the number of iterations by a factor of 10. When performing TE(target encoding), we can use 10 internally nested folds instead of 5. I didn't try it because it would have taken too long.\n

","metadata":{}},{"cell_type":"markdown","source":"# cuDF","metadata":{}},{"cell_type":"markdown","source":"### GPUT4×2 is needed to run it","metadata":{}},{"cell_type":"markdown","source":"

\n 这里我使用了 @cdeotte 发布的 [cuDF]\n

","metadata":{}},{"cell_type":"code","source":"%load_ext cudf.pandas","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-10T02:03:07.178779Z","iopub.execute_input":"2025-01-10T02:03:07.179147Z","iopub.status.idle":"2025-01-10T02:03:07.183725Z","shell.execute_reply.started":"2025-01-10T02:03:07.179115Z","shell.execute_reply":"2025-01-10T02:03:07.182786Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# Date-time processing","metadata":{}},{"cell_type":"code","source":"import numpy as np, pandas as pd\npd.set_option('display.max_columns', 500)\n\ntrain = pd.read_csv(\"/kaggle/input/playground-series-s4e12/train.csv\")\ntrain[\"Policy Start Date\"] = pd.to_datetime( train[\"Policy Start Date\"] )\ntrain[\"year\"] = train[\"Policy Start Date\"].dt.year.astype(\"float32\")\ntrain[\"month\"] = train[\"Policy Start Date\"].dt.month.astype(\"float32\")\ntrain[\"day\"] = train[\"Policy Start Date\"].dt.day.astype(\"float32\")\ntrain[\"dow\"] = train[\"Policy Start Date\"].dt.dayofweek.astype(\"float32\")\ntrain[\"seconds\"] = (train[\"Policy Start Date\"].astype(\"int64\") // 10**9).astype(\"float32\")\n\ntrain[\"y\"] = np.log1p( train[\"Premium Amount\"] )\nprint( train.shape )","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-10T02:03:07.198756Z","iopub.execute_input":"2025-01-10T02:03:07.198957Z","iopub.status.idle":"2025-01-10T02:03:07.464434Z","shell.execute_reply.started":"2025-01-10T02:03:07.198932Z","shell.execute_reply":"2025-01-10T02:03:07.463639Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"train.head()","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"code","source":"test = pd.read_csv(\"/kaggle/input/playground-series-s4e12/test.csv\")\ntest[\"Policy Start Date\"] = pd.to_datetime( test[\"Policy Start Date\"] )\ntest[\"year\"] = test[\"Policy Start Date\"].dt.year.astype(\"float32\")\ntest[\"month\"] = test[\"Policy Start Date\"].dt.month.astype(\"float32\")\ntest[\"day\"] = test[\"Policy Start Date\"].dt.day.astype(\"float32\")\ntest[\"dow\"] = test[\"Policy Start Date\"].dt.dayofweek.astype(\"float32\")\ntest[\"seconds\"] = (test[\"Policy Start Date\"].astype(\"int64\") // 10**9).astype(\"float32\")\nprint( test.shape )\n","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-10T02:03:07.465744Z","iopub.execute_input":"2025-01-10T02:03:07.466066Z","iopub.status.idle":"2025-01-10T02:03:07.653093Z","shell.execute_reply.started":"2025-01-10T02:03:07.466016Z","shell.execute_reply":"2025-01-10T02:03:07.652247Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"test.head()","metadata":{"trusted":true},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# Numerical feature processing","metadata":{}},{"cell_type":"markdown","source":"

\n I applied FOLDS = 20 when I ran all the steps.\n

\n

\n When this treatment is not used, Overall CV RMSLE = 1.01931\n

\n

\n When this method was used, the Overall CV RMSLE = 1.01852\n

\n

\n This is a treatment I explored during the game, which I post in [this discussion], and is helpful for scoring\n

\n

\n Wow, this is really great.\n

","metadata":{}},{"cell_type":"code","source":"\ndef log_transform_features(df, num_cols):\n X_num = df[num_cols]\n X_num_log = pd.DataFrame(np.log1p(X_num), columns=X_num.columns)\n df[num_cols] = X_num_log\n return df\n \nlog_transform_cols = ['Annual Income', 'Previous Claims']\n\n\ntrain = log_transform_features(train, log_transform_cols)\ntest = log_transform_features(test, log_transform_cols)\n","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-10T02:03:07.653988Z","iopub.execute_input":"2025-01-10T02:03:07.654323Z","iopub.status.idle":"2025-01-10T02:03:07.85721Z","shell.execute_reply.started":"2025-01-10T02:03:07.65429Z","shell.execute_reply":"2025-01-10T02:03:07.856533Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# classification","metadata":{}},{"cell_type":"code","source":"RMV = [\"id\",\"Policy Start Date\",\"Premium Amount\",\"y\"]\nFEATURES = [c for c in train.columns if not c in RMV]\ncombined = pd.concat([train,test],axis=0,ignore_index=True)\n\nCATS = []\nHIGH_CARDINALITY = []\nprint(f\"THE {len(FEATURES)} BASIC FEATURES ARE:\")\n\nfor c in FEATURES:\n ftype = \"numerical\"\n if combined[c].dtype==\"object\":\n CATS.append(c)\n combined[c] = combined[c].fillna(\"NAN\")\n combined[c],_ = combined[c].factorize()\n combined[c] -= combined[c].min()\n ftype = \"categorical\"\n if combined[c].dtype==\"int64\":\n combined[c] = combined[c].astype(\"int32\")\n elif combined[c].dtype==\"float64\":\n combined[c] = combined[c].astype(\"float32\")\n \n n = combined[c].nunique()\n print(f\"{c} ({ftype}) with {n} unique values\")\n if n>=9: HIGH_CARDINALITY.append(c)\n \ntrain = combined.iloc[:len(train)].copy()\ntest = combined.iloc[len(train):].reset_index(drop=True).copy()\n\nprint(\"\\nTHE FOLLOWING HAVE 9 OR MORE UNIQUE VALUES:\", HIGH_CARDINALITY )","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-10T02:03:07.857801Z","iopub.execute_input":"2025-01-10T02:03:07.858043Z","iopub.status.idle":"2025-01-10T02:03:08.312983Z","shell.execute_reply.started":"2025-01-10T02:03:07.858024Z","shell.execute_reply":"2025-01-10T02:03:08.31214Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# New feature","metadata":{}},{"cell_type":"code","source":"lists2 = [['Annual Income', 'Health Score'], ['Credit Score', 'Health Score'], \n ['Customer Feedback', 'Gender', 'Marital Status', 'Occupation', 'Smoking Status', 'year'], \n ['Exercise Frequency', 'Health Score'], ['Health Score', 'Marital Status'], \n ['Education Level', 'Gender', 'Health Score'], ['Health Score', 'Occupation'], \n ['Age', 'Health Score'], ['Health Score', 'dow'], ['Age', 'Exercise Frequency', 'Location'], \n ['Health Score', 'Smoking Status', 'month'], ['Health Score', 'Location', 'Policy Type'], \n ['Health Score', 'Insurance Duration'], ['Health Score', 'Number of Dependents'], \n ['Customer Feedback', 'Exercise Frequency', 'Previous Claims', 'Property Type', 'dow'], \n ['Customer Feedback', 'Health Score'], ['Health Score', 'Property Type'], \n ['Health Score', 'day', 'seconds'], ['Health Score', 'year'], ['Age', 'Gender', 'Insurance Duration', 'year']]\nprint(f\"We have {len(lists2)} powerful combination of columns!\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-10T02:03:08.313817Z","iopub.execute_input":"2025-01-10T02:03:08.314133Z","iopub.status.idle":"2025-01-10T02:03:08.319733Z","shell.execute_reply.started":"2025-01-10T02:03:08.314108Z","shell.execute_reply":"2025-01-10T02:03:08.318988Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# Target encode","metadata":{}},{"cell_type":"code","source":"def target_encode(train, valid, test, col, target=\"y\", kfold=5, smooth=20, agg=\"mean\"):\n\n train['kfold'] = ((train.index) % kfold)\n col_name = '_'.join(col)\n train[f'TE_{agg.upper()}_' + col_name] = 0.\n for i in range(kfold):\n \n df_tmp = train[train['kfold']!=i]\n if agg==\"mean\": mn = train[target].mean()\n elif agg==\"median\": mn = train[target].median()\n elif agg==\"min\": mn = train[target].min()\n elif agg==\"max\": mn = train[target].max()\n elif agg==\"nunique\": mn = 0\n df_tmp = df_tmp[col + [target]].groupby(col).agg([agg, 'count']).reset_index()\n df_tmp.columns = col + [agg, 'count']\n if agg==\"nunique\":\n df_tmp['TE_tmp'] = df_tmp[agg] / df_tmp['count']\n else:\n df_tmp['TE_tmp'] = ((df_tmp[agg]*df_tmp['count'])+(mn*smooth)) / (df_tmp['count']+smooth)\n df_tmp_m = train[col + ['kfold', f'TE_{agg.upper()}_' + col_name]].merge(df_tmp, how='left', left_on=col, right_on=col)\n df_tmp_m.loc[df_tmp_m['kfold']==i, f'TE_{agg.upper()}_' + col_name] = df_tmp_m.loc[df_tmp_m['kfold']==i, 'TE_tmp']\n train[f'TE_{agg.upper()}_' + col_name] = df_tmp_m[f'TE_{agg.upper()}_' + col_name].fillna(mn).values \n \n df_tmp = train[col + [target]].groupby(col).agg([agg, 'count']).reset_index()\n if agg==\"mean\": mn = train[target].mean()\n elif agg==\"median\": mn = train[target].median()\n elif agg==\"min\": mn = train[target].min()\n elif agg==\"max\": mn = train[target].max()\n elif agg==\"nunique\": mn = 0\n df_tmp.columns = col + [agg, 'count']\n if agg==\"nunique\":\n df_tmp['TE_tmp'] = df_tmp[agg] / df_tmp['count']\n else:\n df_tmp['TE_tmp'] = ((df_tmp[agg]*df_tmp['count'])+(mn*smooth)) / (df_tmp['count']+smooth)\n df_tmp_m = valid[col].merge(df_tmp, how='left', left_on=col, right_on=col)\n valid[f'TE_{agg.upper()}_' + col_name] = df_tmp_m['TE_tmp'].fillna(mn).values\n valid[f'TE_{agg.upper()}_' + col_name] = valid[f'TE_{agg.upper()}_' + col_name].astype(\"float32\")\n\n df_tmp_m = test[col].merge(df_tmp, how='left', left_on=col, right_on=col)\n test[f'TE_{agg.upper()}_' + col_name] = df_tmp_m['TE_tmp'].fillna(mn).values\n test[f'TE_{agg.upper()}_' + col_name] = test[f'TE_{agg.upper()}_' + col_name].astype(\"float32\")\n\n train = train.drop('kfold', axis=1)\n train[f'TE_{agg.upper()}_' + col_name] = train[f'TE_{agg.upper()}_' + col_name].astype(\"float32\")\n\n return(train, valid, test)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-10T02:03:08.320574Z","iopub.execute_input":"2025-01-10T02:03:08.320857Z","iopub.status.idle":"2025-01-10T02:03:08.337159Z","shell.execute_reply.started":"2025-01-10T02:03:08.320828Z","shell.execute_reply":"2025-01-10T02:03:08.336333Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"train.head()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-10T02:03:08.339434Z","iopub.execute_input":"2025-01-10T02:03:08.339651Z","iopub.status.idle":"2025-01-10T02:03:08.392944Z","shell.execute_reply.started":"2025-01-10T02:03:08.339632Z","shell.execute_reply":"2025-01-10T02:03:08.392086Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"test.head()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-10T02:03:08.394172Z","iopub.execute_input":"2025-01-10T02:03:08.394457Z","iopub.status.idle":"2025-01-10T02:03:08.442891Z","shell.execute_reply.started":"2025-01-10T02:03:08.394429Z","shell.execute_reply":"2025-01-10T02:03:08.442264Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"train.info()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-10T02:03:08.443674Z","iopub.execute_input":"2025-01-10T02:03:08.443896Z","iopub.status.idle":"2025-01-10T02:03:08.462727Z","shell.execute_reply.started":"2025-01-10T02:03:08.443865Z","shell.execute_reply":"2025-01-10T02:03:08.462089Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# XGB","metadata":{}},{"cell_type":"markdown","source":"### If you want to try it, please drop the \"\"\" after it.","metadata":{}},{"cell_type":"code","source":"from xgboost import XGBRegressor\nimport xgboost as xgb, time\nprint(f\"Using XGBoost version\",xgb.__version__)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-10T02:03:08.463414Z","iopub.execute_input":"2025-01-10T02:03:08.463613Z","iopub.status.idle":"2025-01-10T02:03:08.468251Z","shell.execute_reply.started":"2025-01-10T02:03:08.463595Z","shell.execute_reply":"2025-01-10T02:03:08.467436Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"

\n Here, I removed the newly created features from lists2 and processed 23 columns of features to observe the CV score gap between them.\n

\n

\n When I applied FOLDS = 2, the scores for all 20 features were the same with or without Folds = 1.02840. Does not reflect the effect of the 20 new features added.\n

\n

\n So when I used FOLDS = 20, I was able to reflect the effects of the 20 new features.\n

\n

\n When using 20 new features, Overall CV RMSLE = 1.01931.\n

\n

\n When the 20 new features are not used, Overall CV RMSLE = 1.02279.\n

","metadata":{}},{"cell_type":"code","source":"\"\"\"\n%%time\n\n#You can set FOLDS = 20 or FOLDS = 2\n\n#FOLDS = 20\nFOLDS = 2\nfrom sklearn.model_selection import KFold\nkf = KFold(n_splits=FOLDS, shuffle=True, random_state=42)\n\noof = np.zeros(len(train))\npred = np.zeros(len(test))\n\nfor i, (train_index, test_index) in enumerate(kf.split(train)):\n\n print(\"#\"*25)\n print(f\"### Fold {i+1}\")\n print(\"#\"*25)\n \n x_train = train.loc[train_index,FEATURES+[\"y\"] ].copy()\n y_train = train.loc[train_index,\"y\"]\n x_valid = train.loc[test_index,FEATURES].copy()\n y_valid = train.loc[test_index,\"y\"]\n x_test = test[FEATURES].copy()\n\n start = time.time()\n print(f\"FEATURE ENGINEER {len(FEATURES)} COLUMNS: \",end=\"\")\n for j,f in enumerate(FEATURES):\n\n c = [f]\n print(f\"({j+1}){c}\",\", \",end=\"\")\n\n # LOW CARDINALITY FEATURES - TARGET ENCODE MEAN AND MEDIAN\n x_train, x_valid, x_test = target_encode(x_train, x_valid, x_test, c, smooth=20, agg=\"mean\")\n x_train, x_valid, x_test = target_encode(x_train, x_valid, x_test, c, smooth=0, agg=\"median\")\n\n # HIGH CARDINALITY FEATURES - TE MIN, MAX, NUNIQUE and CE\n if c[0] in HIGH_CARDINALITY:\n x_train, x_valid, x_test = target_encode(x_train, x_valid, x_test, c, smooth=0, agg=\"min\")\n x_train, x_valid, x_test = target_encode(x_train, x_valid, x_test, c, smooth=0, agg=\"max\")\n x_train, x_valid, x_test = target_encode(x_train, x_valid, x_test, c, smooth=0, agg=\"nunique\")\n \n # COUNT ENCODING (USING COMBINED TRAIN TEST)\n tmp = combined.groupby(c).y.count()\n nm = f\"CE_{'_'.join(c)}\"; tmp.name = nm\n x_train = x_train.merge(tmp, on=c, how=\"left\")\n x_valid = x_valid.merge(tmp, on=c, how=\"left\")\n x_test = x_test.merge(tmp, on=c, how=\"left\")\n x_train[nm] = x_train[nm].astype(\"int32\")\n x_valid[nm] = x_valid[nm].astype(\"int32\")\n x_test[nm] = x_test[nm].astype(\"int32\")\n \n end = time.time()\n elapsed = end-start\n print(f\"Feature engineering took {elapsed:.1f} seconds\")\n x_train = x_train.drop(\"y\",axis=1)\n\n model = XGBRegressor(\n device=\"cuda\",\n max_depth=8, \n colsample_bytree=0.9, \n subsample=0.9, \n n_estimators=2_000, \n learning_rate=0.01, \n early_stopping_rounds=25, \n eval_metric=\"rmse\",\n )\n model.fit(\n x_train, y_train,\n eval_set=[(x_valid, y_valid)], \n verbose=100\n )\n\n # INFER OOF\n oof[test_index] = model.predict(x_valid)\n # INFER TEST\n pred += model.predict(x_test)\n\n m = np.sqrt(np.mean( (y_valid.to_numpy() - oof[test_index])**2.0 )) \n print(f\" => Fold {i+1} RMSLE = {m:.5f}\")\n\n# COMPUTE AVERAGE TEST PREDS\npred /= FOLDS\n\"\"\"","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-10T02:03:08.469143Z","iopub.execute_input":"2025-01-10T02:03:08.469464Z","iopub.status.idle":"2025-01-10T02:03:08.482494Z","shell.execute_reply.started":"2025-01-10T02:03:08.46943Z","shell.execute_reply":"2025-01-10T02:03:08.481691Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"在这里我将FOLDS = 20修改为了FOLDS = 2.\n\n当FOLDS=20时，Overall CV RMSLE = 1.01931\n\n当FOLDS=2时，Overall CV RMSLE = 1.02840\n对'Annual Income', 'Previous Claims'进行log转化，Overall CV RMSLE = 1.02457","metadata":{}},{"cell_type":"code","source":"\"\"\"\n%%time\n#FOLDS = 20\nFOLDS = 2\nfrom sklearn.model_selection import KFold\nkf = KFold(n_splits=FOLDS, shuffle=True, random_state=42)\n\noof = np.zeros(len(train))\npred = np.zeros(len(test))\n\nfor i, (train_index, test_index) in enumerate(kf.split(train)):\n\n print(\"#\"*25)\n print(f\"### Fold {i+1}\")\n print(\"#\"*25)\n \n x_train = train.loc[train_index,FEATURES+[\"y\"] ].copy()\n y_train = train.loc[train_index,\"y\"]\n x_valid = train.loc[test_index,FEATURES].copy()\n y_valid = train.loc[test_index,\"y\"]\n x_test = test[FEATURES].copy()\n\n start = time.time()\n print(f\"FEATURE ENGINEER {len(FEATURES)} COLUMNS and {len(lists2)} GROUPS: \",end=\"\")\n for j,f in enumerate(FEATURES+lists2):\n\n if j=len(FEATURES)) | (c[0] in HIGH_CARDINALITY):\n x_train, x_valid, x_test = target_encode(x_train, x_valid, x_test, c, smooth=0, agg=\"min\")\n x_train, x_valid, x_test = target_encode(x_train, x_valid, x_test, c, smooth=0, agg=\"max\")\n x_train, x_valid, x_test = target_encode(x_train, x_valid, x_test, c, smooth=0, agg=\"nunique\")\n \n # COUNT ENCODING (USING COMBINED TRAIN TEST)\n tmp = combined.groupby(c).y.count()\n nm = f\"CE_{'_'.join(c)}\"; tmp.name = nm\n x_train = x_train.merge(tmp, on=c, how=\"left\")\n x_valid = x_valid.merge(tmp, on=c, how=\"left\")\n x_test = x_test.merge(tmp, on=c, how=\"left\")\n x_train[nm] = x_train[nm].astype(\"int32\")\n x_valid[nm] = x_valid[nm].astype(\"int32\")\n x_test[nm] = x_test[nm].astype(\"int32\")\n \n end = time.time()\n elapsed = end-start\n print(f\"Feature engineering took {elapsed:.1f} seconds\")\n x_train = x_train.drop(\"y\",axis=1)\n\n model = XGBRegressor(\n device=\"cuda\",\n max_depth=8, \n colsample_bytree=0.9, \n subsample=0.9, \n n_estimators=2_000, \n learning_rate=0.01, \n early_stopping_rounds=25, \n eval_metric=\"rmse\",\n )\n model.fit(\n x_train, y_train,\n eval_set=[(x_valid, y_valid)], \n verbose=100\n )\n\n # INFER OOF\n oof[test_index] = model.predict(x_valid)\n # INFER TEST\n pred += model.predict(x_test)\n\n m = np.sqrt(np.mean( (y_valid.to_numpy() - oof[test_index])**2.0 )) \n print(f\" => Fold {i+1} RMSLE = {m:.5f}\")\n\n# COMPUTE AVERAGE TEST PREDS\npred /= FOLDS\n\"\"\"","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-10T02:03:08.483398Z","iopub.execute_input":"2025-01-10T02:03:08.483689Z","iopub.status.idle":"2025-01-10T03:46:43.569163Z","shell.execute_reply.started":"2025-01-10T02:03:08.483661Z","shell.execute_reply":"2025-01-10T03:46:43.568212Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# Verification score","metadata":{}},{"cell_type":"code","source":"\"\"\"\nm = np.sqrt(np.mean( (train.y.values - oof)**2.0 )) \nprint(f\"Overall CV RMSLE = {m:.5f}\")\n\"\"\"","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-10T03:46:43.569874Z","iopub.execute_input":"2025-01-10T03:46:43.570117Z","iopub.status.idle":"2025-01-10T03:46:43.588362Z","shell.execute_reply.started":"2025-01-10T03:46:43.570097Z","shell.execute_reply":"2025-01-10T03:46:43.587686Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"# Conclusion","metadata":{}},{"cell_type":"markdown","source":"

\n I learned that the gap with the first place has these aspects:\n

\n
    \n
  1. He added a new feature list
    2. \n
  2. He uses TE(object coding) for features, which is a great way to do it, and of course, the effect is very good.
    3. \n
  3. His verification method fits perfectly
    4. \n
","metadata":{}}]}