{"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":"none","dataSources":[{"sourceId":84896,"databundleVersionId":10305135,"sourceType":"competition"}],"dockerImageVersionId":30822,"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":false}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"# This Python 3 environment comes with many helpful analytics libraries installed\n# It is defined by the kaggle/python Docker image: https://github.com/kaggle/docker-python\n# For example, here's several helpful packages to load\n\nimport numpy as np # linear algebra\nimport pandas as pd # data processing, CSV file I/O (e.g. pd.read_csv)\n\n# Input data files are available in the read-only \"../input/\" directory\n# For example, running this (by clicking run or pressing Shift+Enter) will list all files under the input directory\n\nimport os\nfor dirname, _, filenames in os.walk('/kaggle/input'):\n    for filename in filenames:\n        print(os.path.join(dirname, filename))\n\n# You can write up to 20GB to the current directory (/kaggle/working/) that gets preserved as output when you create a version using \"Save & Run All\" \n# You can also write temporary files to /kaggle/temp/, but they won't be saved outside of the current session","metadata":{"_uuid":"8f2839f25d086af736a60e9eeb907d3b93b6e0e5","_cell_guid":"b1076dfc-b9ad-4769-8c92-a6c4dae69d19","trusted":true,"execution":{"iopub.status.busy":"2024-12-26T13:14:08.259668Z","iopub.execute_input":"2024-12-26T13:14:08.260022Z","iopub.status.idle":"2024-12-26T13:14:08.660519Z","shell.execute_reply.started":"2024-12-26T13:14:08.259991Z","shell.execute_reply":"2024-12-26T13:14:08.659483Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"import pandas as pd\nfrom lightgbm import LGBMRegressor\nfrom sklearn.model_selection import KFold\nfrom sklearn.metrics import mean_squared_error, mean_squared_log_error\nfrom sklearn.preprocessing import LabelEncoder, StandardScaler","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-26T13:14:11.36711Z","iopub.execute_input":"2024-12-26T13:14:11.367466Z","iopub.status.idle":"2024-12-26T13:14:14.2038Z","shell.execute_reply.started":"2024-12-26T13:14:11.367437Z","shell.execute_reply":"2024-12-26T13:14:14.202859Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"train = pd.read_csv('/kaggle/input/playground-series-s4e12/train.csv')\ntest = pd.read_csv('/kaggle/input/playground-series-s4e12/test.csv')","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-26T13:14:14.20515Z","iopub.execute_input":"2024-12-26T13:14:14.205941Z","iopub.status.idle":"2024-12-26T13:14:24.48455Z","shell.execute_reply.started":"2024-12-26T13:14:14.205903Z","shell.execute_reply":"2024-12-26T13:14:24.483635Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"print(train.shape)\nprint(test.shape)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-26T13:14:24.486067Z","iopub.execute_input":"2024-12-26T13:14:24.486357Z","iopub.status.idle":"2024-12-26T13:14:24.491062Z","shell.execute_reply.started":"2024-12-26T13:14:24.486326Z","shell.execute_reply":"2024-12-26T13:14:24.490122Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"train.head()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-26T13:14:24.492715Z","iopub.execute_input":"2024-12-26T13:14:24.492995Z","iopub.status.idle":"2024-12-26T13:14:24.54786Z","shell.execute_reply.started":"2024-12-26T13:14:24.492963Z","shell.execute_reply":"2024-12-26T13:14:24.546622Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"train.isnull().sum()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-26T13:14:24.548864Z","iopub.execute_input":"2024-12-26T13:14:24.549191Z","iopub.status.idle":"2024-12-26T13:14:25.170695Z","shell.execute_reply.started":"2024-12-26T13:14:24.549153Z","shell.execute_reply":"2024-12-26T13:14:25.169726Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"train['Health Score'].dropna().astype(int).nunique()\ntest['Health Score'].dropna().astype(int).nunique()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-26T13:14:25.171544Z","iopub.execute_input":"2024-12-26T13:14:25.171904Z","iopub.status.idle":"2024-12-26T13:14:25.213093Z","shell.execute_reply.started":"2024-12-26T13:14:25.171877Z","shell.execute_reply":"2024-12-26T13:14:25.212035Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"target = train['Premium Amount']\n\nnumeric_columns = train.select_dtypes(exclude=['object']).columns.tolist()\nnumeric_columns.remove('id')\nnumeric_columns.remove('Premium Amount')\n\nscaler = StandardScaler()\ntrain[numeric_columns] = scaler.fit_transform(train[numeric_columns])\ntest[numeric_columns] = scaler.fit_transform(test[numeric_columns])\n\ntrain['Premium Amount'] = np.log1p(train['Premium Amount'])","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-26T13:14:25.214041Z","iopub.execute_input":"2024-12-26T13:14:25.214366Z","iopub.status.idle":"2024-12-26T13:14:25.865939Z","shell.execute_reply.started":"2024-12-26T13:14:25.214322Z","shell.execute_reply":"2024-12-26T13:14:25.864907Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# replace missing values\nfor df in [train, test]:\n    df.fillna({\n        'Age': train['Age'].mean(),\n        'Annual Income': train['Annual Income'].mean(),\n        'Number of Dependents': train['Number of Dependents'].mean(),\n        'Health Score': train['Health Score'].mean(),\n        'Previous Claims': train['Previous Claims'].mean(),\n        'Vehicle Age': train['Vehicle Age'].mean(),\n        'Credit Score': train['Credit Score'].mean(),\n        'Insurance Duration': train['Insurance Duration'].mean(),\n        'Marital Status': 'Unknown',\n        'Occupation': 'Unknown',\n        'Customer Feedback': 'Unknown',\n    }, inplace=True)\n\n# Split Policy Start Date into day, date, month and year\nfor df in [train, test]:\n    policy_start_date = pd.to_datetime(df['Policy Start Date'])\n    df['Year'] = policy_start_date.dt.year\n    df['Month'] = policy_start_date.dt.month\n    df['Day'] = policy_start_date.dt.day\n    df['Year_sin'] = np.sin(2 * np.pi * df['Year'])\n    df.drop('Policy Start Date',axis=1,inplace=True)\n\n# convert columns to category type wherever possible\ncategorical_columns  = train.select_dtypes(include = \"object\").columns\nfor df in [train, test]:\n    for col in categorical_columns:\n        df[col] = df[col].astype('category')\n    df = pd.get_dummies(df, columns=categorical_columns)\n\ncombined = pd.concat([train, test], axis=0, ignore_index=True)\n\nfor df in [train, test]:\n    for col in categorical_columns:\n        freq_encoding = combined[col].value_counts().to_dict()\n        df[f\"{col}_freq\"] = df[col].map(freq_encoding).astype('float')\n        df.drop(col,axis=1,inplace=True)\n","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-26T13:14:25.86884Z","iopub.execute_input":"2024-12-26T13:14:25.869137Z","iopub.status.idle":"2024-12-26T13:14:31.146091Z","shell.execute_reply.started":"2024-12-26T13:14:25.869111Z","shell.execute_reply":"2024-12-26T13:14:31.14502Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"def rmsle(y_true, y_pred):\n    y_pred = np.maximum(0, y_pred)  # Clip predicted values to be non-negative\n    return np.sqrt(mean_squared_log_error(y_true, y_pred))","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-26T13:14:31.147216Z","iopub.execute_input":"2024-12-26T13:14:31.147474Z","iopub.status.idle":"2024-12-26T13:14:31.151903Z","shell.execute_reply.started":"2024-12-26T13:14:31.147452Z","shell.execute_reply":"2024-12-26T13:14:31.150811Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"X = train.drop(columns=['id', 'Premium Amount'])\ny_log = train['Premium Amount']\n\nlgb_params = {\n    'num_leaves': 71,\n    'learning_rate': 0.05412467152424433,\n    'n_estimators': 595,\n    'max_depth': 12,\n    'min_data_in_leaf': 97,\n    'bagging_fraction': 0.5200288825838669,\n    'feature_fraction': 0.9881738491942492,\n    'n_jobs': -1,\n    'verbose': -1\n}\n\ndef train_model():\n    kf = KFold(n_splits=10, shuffle=True, random_state=42)\n    oof = np.zeros(len(X))\n    models = []\n\n    for fold, (train_idx, valid_idx) in enumerate(kf.split(X)):\n        print(f\"Fold {fold + 1}\")\n        X_train, X_valid = X.iloc[train_idx], X.iloc[valid_idx]\n        y_train, y_valid = y_log.iloc[train_idx], y_log.iloc[valid_idx]\n\n        lgbm_model = LGBMRegressor(**lgb_params)\n\n        lgbm_model.fit(X_train, y_train)\n        oof[valid_idx] = np.maximum(0, lgbm_model.predict(X_valid))\n        fold_rmsle = rmsle(np.expm1(y_valid), np.expm1(oof[valid_idx]))\n        print(f\"Fold {fold + 1} RMSLE: {fold_rmsle}\")\n        models.append(lgbm_model)\n\n    return models, oof","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-26T13:14:31.15301Z","iopub.execute_input":"2024-12-26T13:14:31.15342Z","iopub.status.idle":"2024-12-26T13:14:31.266878Z","shell.execute_reply.started":"2024-12-26T13:14:31.153383Z","shell.execute_reply":"2024-12-26T13:14:31.265947Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"models,oof = train_model()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-26T13:14:31.267784Z","iopub.execute_input":"2024-12-26T13:14:31.26806Z","iopub.status.idle":"2024-12-26T13:19:08.978769Z","shell.execute_reply.started":"2024-12-26T13:14:31.268037Z","shell.execute_reply":"2024-12-26T13:19:08.977703Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"print(rmsle(np.expm1(train['Premium Amount']), np.expm1(oof)))\n# 1.0457401258689165","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-26T13:19:08.979783Z","iopub.execute_input":"2024-12-26T13:19:08.980113Z","iopub.status.idle":"2024-12-26T13:19:09.092256Z","shell.execute_reply.started":"2024-12-26T13:19:08.980075Z","shell.execute_reply":"2024-12-26T13:19:09.091191Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"X_test = test.drop(columns=['id'])\ntest_predictions = np.zeros(len(test))\n\nfor model in models:\n    test_predictions += np.maximum(0, np.expm1(model.predict(X_test))) / len(models)\n\nsubmission = test[['id']].copy()\nsubmission['Premium Amount'] = test_predictions\nsubmission.to_csv('submission.csv', index = False)\nsubmission.head()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2024-12-26T13:19:09.093198Z","iopub.execute_input":"2024-12-26T13:19:09.093579Z","iopub.status.idle":"2024-12-26T13:20:38.614054Z","shell.execute_reply.started":"2024-12-26T13:19:09.093544Z","shell.execute_reply":"2024-12-26T13:20:38.612956Z"}},"outputs":[],"execution_count":null}]}