{"metadata":{"kernelspec":{"name":"python3","display_name":"Python 3","language":"python"},"language_info":{"name":"python","version":"3.10.12","mimetype":"text/x-python","codemirror_mode":{"name":"ipython","version":3},"pygments_lexer":"ipython3","nbconvert_exporter":"python","file_extension":".py"},"kaggle":{"accelerator":"gpu","dataSources":[{"sourceId":84896,"databundleVersionId":10305135,"sourceType":"competition"}],"dockerImageVersionId":30823,"isInternetEnabled":true,"language":"python","sourceType":"notebook","isGpuEnabled":true}},"nbformat_minor":4,"nbformat":4,"cells":[{"cell_type":"code","source":"import warnings\nwarnings.filterwarnings('ignore')","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T05:58:33.211797Z","iopub.execute_input":"2025-01-02T05:58:33.212072Z","iopub.status.idle":"2025-01-02T05:58:33.215821Z","shell.execute_reply.started":"2025-01-02T05:58:33.212043Z","shell.execute_reply":"2025-01-02T05:58:33.215133Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"import pandas as pd\ndata_path=\"/kaggle/input/playground-series-s4e12/\"\ntrain=pd.read_csv(data_path+'train.csv')\ntest=pd.read_csv(data_path+'test.csv')\nsubmission=pd.read_csv(data_path+'sample_submission.csv')\n\ntrain=train.drop('id',axis=1)\ntest=test.drop('id',axis=1)\n\ntrain.shape, test.shape, submission.shape","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T05:58:33.21635Z","iopub.execute_input":"2025-01-02T05:58:33.216617Z","iopub.status.idle":"2025-01-02T05:58:42.326346Z","shell.execute_reply.started":"2025-01-02T05:58:33.216591Z","shell.execute_reply":"2025-01-02T05:58:42.325551Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"train.head()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T05:58:42.328215Z","iopub.execute_input":"2025-01-02T05:58:42.328492Z","iopub.status.idle":"2025-01-02T05:58:42.351598Z","shell.execute_reply.started":"2025-01-02T05:58:42.328471Z","shell.execute_reply":"2025-01-02T05:58:42.350833Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"test.head()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T05:58:42.352797Z","iopub.execute_input":"2025-01-02T05:58:42.353028Z","iopub.status.idle":"2025-01-02T05:58:42.369108Z","shell.execute_reply.started":"2025-01-02T05:58:42.353008Z","shell.execute_reply":"2025-01-02T05:58:42.368344Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"submission.head()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T05:58:42.36984Z","iopub.execute_input":"2025-01-02T05:58:42.370093Z","iopub.status.idle":"2025-01-02T05:58:42.387138Z","shell.execute_reply.started":"2025-01-02T05:58:42.370073Z","shell.execute_reply":"2025-01-02T05:58:42.386442Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"submission은 test의 id에 대해서 모든 Premium Amount가 1102.545로 고정되어있다.<br><br>\n\n우리가 구해야하는 값은 Premium Amount이다. ","metadata":{}},{"cell_type":"code","source":"def resumetable(df):\n    print(f'데이터 세트 형상: {df.shape}')\n    summary = pd.DataFrame(df.dtypes, columns=['데이터 타입'])\n    summary = summary.reset_index()\n    summary = summary.rename(columns={'index': '피처'})\n    summary['결측값 개수'] = df.isnull().sum().values\n    summary['고윳값 개수'] = df.nunique().values\n    summary['첫 번째 값'] = df.loc[0].values\n    summary['두 번째 값'] = df.loc[1].values\n    summary['세 번째 값'] = df.loc[2].values\n    summary['? 값'] = df.loc[453].values\n\n    return summary\n    \nresumetable(train)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T05:58:42.388001Z","iopub.execute_input":"2025-01-02T05:58:42.388296Z","iopub.status.idle":"2025-01-02T05:58:43.777421Z","shell.execute_reply.started":"2025-01-02T05:58:42.388267Z","shell.execute_reply":"2025-01-02T05:58:43.77658Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"resumetable(test)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T05:58:43.77821Z","iopub.execute_input":"2025-01-02T05:58:43.778472Z","iopub.status.idle":"2025-01-02T05:58:44.684302Z","shell.execute_reply.started":"2025-01-02T05:58:43.778447Z","shell.execute_reply":"2025-01-02T05:58:44.683543Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"Gender, Marital Status(결혼 상태), Number of Dependents(부양 가족 수), Education Level(교육 수준), Occupation(직업), Policy Type(정책 형태), Previous Claims(이전 청구), Vehicle Age(차량 연식), Credit Score(신용 점수), Insurance Duration(보험 기간), Customer Feedback(고객 피드백), Smoking Status, Exercise Frequency(운동 빈도), Property Type(재산 유형)","metadata":{}},{"cell_type":"code","source":"object_columns=train.select_dtypes(include=['object']).columns","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T05:58:44.685073Z","iopub.execute_input":"2025-01-02T05:58:44.685388Z","iopub.status.idle":"2025-01-02T05:58:44.830957Z","shell.execute_reply.started":"2025-01-02T05:58:44.685343Z","shell.execute_reply":"2025-01-02T05:58:44.829929Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"for col in object_columns:\n    print(col, train[col].unique())","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T05:58:44.833648Z","iopub.execute_input":"2025-01-02T05:58:44.833883Z","iopub.status.idle":"2025-01-02T05:58:45.524172Z","shell.execute_reply.started":"2025-01-02T05:58:44.833865Z","shell.execute_reply":"2025-01-02T05:58:45.523446Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"Marital Status, Occupation, Customer Feedback 은 결측치가 있어서 따로 처리를 해줘야 한다.<br><br>\n\nMarital Status은 위 순서대로 `결혼, 이혼, 미혼`<br>\nEducation Level은 위 순서대로 `학사, 석사, 고등학교, 박사`<br>\nOccupation은 위 순서대로 `자영업자, 취업자, 실업자`<br>\nLocation은 위 순서대로 `도시, 농촌, 교외`<br>\nPolicy Type은 위 순서대로 `프리미엄, 종합, 기본`<br>\nCustomer Feedback은 위 순서대로 `나쁨, 보통, 좋음`<br>\nExercise Frequency는 위 순서대로 `주간, 월간, 매일, 드물게`<br>\nProperty Type은 위 순서대로 `주택, 아파트, 콘도`<br><br>\n\nEducation Level은 고등학교, 학사, 석사, 박사 순으로 ordinal 타입으로<br>\nPolicy Type은 기본, 종합, 프리미엄 순으로 ordinal 타입으로<br>\nCustomer Feedback은 나쁨, 보통, 좋음 순으로 ordinal 타입으로<br>\nExercise Frequency는 드물게, 월간, 주간, 매일 순으로 ordinal 타입으로<br>\n나머지 데이터는 one-hot encoding 해준다.","metadata":{}},{"cell_type":"code","source":"import seaborn as sns\nimport numpy as np\nsns.displot(train['Premium Amount'])\nsns.displot(np.log1p(train['Premium Amount']))","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T05:58:45.525672Z","iopub.execute_input":"2025-01-02T05:58:45.525879Z","iopub.status.idle":"2025-01-02T05:58:48.799869Z","shell.execute_reply.started":"2025-01-02T05:58:45.525862Z","shell.execute_reply":"2025-01-02T05:58:48.79914Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"타겟인 Premium Amount가 왼쪽으로 치우쳐져 있는 것을 확인할 수 있다. -> 로그 변환을 해준다.(정규 분포를 따른다.)","metadata":{}},{"cell_type":"code","source":"numeric_columns=train.select_dtypes(include=['int64', 'float64']).columns","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T05:58:48.800652Z","iopub.execute_input":"2025-01-02T05:58:48.801087Z","iopub.status.idle":"2025-01-02T05:58:48.836709Z","shell.execute_reply.started":"2025-01-02T05:58:48.801063Z","shell.execute_reply":"2025-01-02T05:58:48.836034Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"pd.set_option('display.float_format', lambda x: '%.2f' % x)\ntrain[numeric_columns].describe()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T05:58:48.837402Z","iopub.execute_input":"2025-01-02T05:58:48.837641Z","iopub.status.idle":"2025-01-02T05:58:49.40983Z","shell.execute_reply.started":"2025-01-02T05:58:48.837621Z","shell.execute_reply":"2025-01-02T05:58:49.408976Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"import matplotlib as mpl\nimport matplotlib.pyplot as plt\nplt.figure(figsize=(12, 10))\nsns.heatmap(train[numeric_columns].corr(), \n            annot=True,\n            cmap='coolwarm',\n            center=0,\n            fmt='.2f',\n            square=True)\nplt.title('수치형 변수 간 상관관계 히트맵', pad=20)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T05:58:49.410619Z","iopub.execute_input":"2025-01-02T05:58:49.410867Z","iopub.status.idle":"2025-01-02T05:58:50.299579Z","shell.execute_reply.started":"2025-01-02T05:58:49.410849Z","shell.execute_reply":"2025-01-02T05:58:50.298763Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"datetime_col=pd.to_datetime(train['Policy Start Date'])\n\ntrain.insert(15, 'Year', datetime_col.dt.year)\ntrain.insert(16, 'Month', datetime_col.dt.month)\ntrain.insert(17, 'Day', datetime_col.dt.day)\ntrain.insert(18, 'Hour', datetime_col.dt.hour)\ntrain.insert(19, 'Minute', datetime_col.dt.minute)\ntrain.insert(20, 'Second', datetime_col.dt.second)\ntrain.insert(21, 'Microsecond', datetime_col.dt.microsecond)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T05:58:50.300601Z","iopub.execute_input":"2025-01-02T05:58:50.300908Z","iopub.status.idle":"2025-01-02T05:58:51.007092Z","shell.execute_reply.started":"2025-01-02T05:58:50.300878Z","shell.execute_reply":"2025-01-02T05:58:51.006376Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"train[['Policy Start Date','Year','Month','Day','Hour','Minute','Second','Microsecond']]","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T05:58:51.007898Z","iopub.execute_input":"2025-01-02T05:58:51.008147Z","iopub.status.idle":"2025-01-02T05:58:51.057906Z","shell.execute_reply.started":"2025-01-02T05:58:51.008128Z","shell.execute_reply":"2025-01-02T05:58:51.057172Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"resumetable(train)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T05:58:51.058628Z","iopub.execute_input":"2025-01-02T05:58:51.058906Z","iopub.status.idle":"2025-01-02T05:58:52.450781Z","shell.execute_reply.started":"2025-01-02T05:58:51.058885Z","shell.execute_reply":"2025-01-02T05:58:52.449896Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# 스텝 1 : m행 n열 Figure 준비\nmpl.rc('font', size=14)       # 폰트 크기 설정\nmpl.rc('axes', titlesize=15)  # 각 축의 제목 크기 설정\nfigure, axes = plt.subplots(nrows=3, ncols=2) # 3행 2열 Figure 생성 \nplt.tight_layout()            # 그래프 사이에 여백 확보 \nfigure.set_size_inches(10, 9) # 전체 Figure 크기를 10x9인치로 설정 \n\n# 스텝 2 : 각 축에 서브플롯 할당\n# 각 축에 연도, 월, 일, 시간, 분, 초별 평균 대여 수량 막대 그래프 할당\nsns.barplot(x='Year', y='Premium Amount', data=train, ax=axes[0,0])\nsns.barplot(x='Month', y='Premium Amount', data=train, ax=axes[0,1])\nsns.barplot(x='Day', y='Premium Amount', data=train, ax=axes[1,0])\nsns.barplot(x='Hour', y='Premium Amount', data=train, ax=axes[1,1])\nsns.barplot(x='Minute', y='Premium Amount', data=train, ax=axes[2,0])\nsns.barplot(x='Second', y='Premium Amount', data=train, ax=axes[2,1])\n\n# 스텝 3 : 세부 설정\n# 3-1 : 서브플롯에 제목 달기\naxes[0,0].set(title='Premium Amount by year')\naxes[0,1].set(title='Premium Amount by month')\naxes[1,0].set(title='Premium Amount by day')\naxes[1,1].set(title='Premium Amount by hour')\naxes[2,0].set(title='Premium Amount by minute')\naxes[2,1].set(title='Premium Amount by second')\n","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T05:58:52.451546Z","iopub.execute_input":"2025-01-02T05:58:52.451766Z","iopub.status.idle":"2025-01-02T05:59:47.980524Z","shell.execute_reply.started":"2025-01-02T05:58:52.451749Z","shell.execute_reply":"2025-01-02T05:59:47.97966Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"값이 하나밖에 없는 `Hour, Minute, Second`는 제거한다.","metadata":{}},{"cell_type":"code","source":"object_columns=object_columns.drop('Policy Start Date')\nprint(object_columns)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T05:59:47.981346Z","iopub.execute_input":"2025-01-02T05:59:47.981594Z","iopub.status.idle":"2025-01-02T05:59:47.986322Z","shell.execute_reply.started":"2025-01-02T05:59:47.981573Z","shell.execute_reply":"2025-01-02T05:59:47.985582Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# object 컬럼 개수에 맞춰 서브플롯 생성\nn_cols = 2  # 한 행에 2개의 그래프\nn_rows = (len(object_columns) + 1) // 2  # 필요한 행 수 계산\n\n# Figure 설정\nmpl.rc('font', size=12)\nmpl.rc('axes', titlesize=13)\nfigure, axes = plt.subplots(nrows=n_rows, ncols=n_cols)\nplt.tight_layout()  # 여백 조정\nfigure.set_size_inches(15, 18)  # Figure 크기 설정\n\n# 각 object 컬럼에 대해 barplot 생성\nfor idx, column in enumerate(object_columns):\n    row = idx // 2\n    col = idx % 2\n    \n    sns.barplot(x=column, y='Premium Amount', data=train, ax=axes[row, col])\n    axes[row, col].set_title(f'Premium Amount by {column}')\n\n# 남는 서브플롯 제거 (object 컬럼 수가 홀수일 경우)\nif len(object_columns) % 2 == 1:\n    figure.delaxes(axes[n_rows-1, 1])\n\nplt.show()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T05:59:47.987065Z","iopub.execute_input":"2025-01-02T05:59:47.98725Z","iopub.status.idle":"2025-01-02T06:01:08.947989Z","shell.execute_reply.started":"2025-01-02T05:59:47.987234Z","shell.execute_reply":"2025-01-02T06:01:08.947136Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"EDA 내용 추가","metadata":{}},{"cell_type":"code","source":"# # 방법 1: 단순 그룹화\n# train.groupby('Marital Status').head(5)\n\n# # 또는 방법 2: 특정 결혼 상태만 보기\n# train[train['Marital Status'] == 'Married'].head(5)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:08.948854Z","iopub.execute_input":"2025-01-02T06:01:08.949218Z","iopub.status.idle":"2025-01-02T06:01:08.953344Z","shell.execute_reply.started":"2025-01-02T06:01:08.94918Z","shell.execute_reply":"2025-01-02T06:01:08.952381Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# train.groupby('Education Level').head(5)\ntrain[train['Age'].isna()].head(20)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:08.954205Z","iopub.execute_input":"2025-01-02T06:01:08.954517Z","iopub.status.idle":"2025-01-02T06:01:09.001856Z","shell.execute_reply.started":"2025-01-02T06:01:08.954478Z","shell.execute_reply":"2025-01-02T06:01:09.000993Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"train[train['Education Level']=='High School']['Age'].mean(),train[train['Education Level']=='Bachelor\\'s']['Age'].mean(),train[train['Education Level']=='Master\\'s']['Age'].mean(),train[train['Education Level']=='PhD']['Age'].mean(),train['Age'].mean()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:09.002749Z","iopub.execute_input":"2025-01-02T06:01:09.003031Z","iopub.status.idle":"2025-01-02T06:01:09.541427Z","shell.execute_reply.started":"2025-01-02T06:01:09.003009Z","shell.execute_reply":"2025-01-02T06:01:09.540538Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"","metadata":{}},{"cell_type":"markdown","source":"### Baseline Model(Linear Regression)","metadata":{}},{"cell_type":"code","source":"train=pd.read_csv(data_path+'train.csv')\ntest=pd.read_csv(data_path+'test.csv')\nsubmission=pd.read_csv(data_path+'sample_submission.csv')\n\ntrain=train.drop('id',axis=1)\ntest=test.drop('id',axis=1)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:09.542388Z","iopub.execute_input":"2025-01-02T06:01:09.542728Z","iopub.status.idle":"2025-01-02T06:01:15.64475Z","shell.execute_reply.started":"2025-01-02T06:01:09.542697Z","shell.execute_reply":"2025-01-02T06:01:15.64405Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"all_data=pd.concat([train, test],ignore_index=True)\nall_data=all_data.drop('Premium Amount', axis=1)\nobject_columns=all_data.select_dtypes(include=['object']).columns\nnumeric_columns=all_data.select_dtypes(include=['int64', 'float64']).columns","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:15.645558Z","iopub.execute_input":"2025-01-02T06:01:15.645763Z","iopub.status.idle":"2025-01-02T06:01:17.212798Z","shell.execute_reply.started":"2025-01-02T06:01:15.645747Z","shell.execute_reply":"2025-01-02T06:01:17.211871Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"from sklearn.linear_model import LinearRegression\nfrom sklearn.model_selection import train_test_split\nfrom sklearn.preprocessing import LabelEncoder, StandardScaler\nfrom sklearn.metrics import mean_squared_error\n\nscaler=StandardScaler()\n# 1. 수치형 데이터 처리\nnumeric_data = all_data[numeric_columns].fillna(0)\nnumeric_data_scaled = pd.DataFrame(\n    scaler.fit_transform(numeric_data),\n    columns=numeric_columns\n)\n\n# 2. 범주형 데이터 처리\ncategorical_data = all_data[object_columns].fillna('None')\n\n# 3. LabelEncoder로 범주형 데이터 변환\nle = LabelEncoder()\nencoded_data = categorical_data.apply(le.fit_transform)\n\n# 4. 모든 데이터 합치기\nall_data_processed = pd.concat([numeric_data_scaled, encoded_data], axis=1)\n\nall_data_processed.head()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:17.213771Z","iopub.execute_input":"2025-01-02T06:01:17.214015Z","iopub.status.idle":"2025-01-02T06:01:23.71192Z","shell.execute_reply.started":"2025-01-02T06:01:17.213995Z","shell.execute_reply":"2025-01-02T06:01:23.710969Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"train_processed=all_data_processed[:len(train)]\ntest_processed=all_data_processed[len(train):]\ntrain_y=np.log1p(train['Premium Amount'])\n\ntrain_x, val_x, train_y, val_y=train_test_split(train_processed, train_y, test_size=0.3, random_state=42)\n\nlr=LinearRegression()\nlr.fit(train_x, train_y)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:23.712856Z","iopub.execute_input":"2025-01-02T06:01:23.713123Z","iopub.status.idle":"2025-01-02T06:01:24.883481Z","shell.execute_reply.started":"2025-01-02T06:01:23.713098Z","shell.execute_reply":"2025-01-02T06:01:24.882051Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"def rmsle(y_true, y_pred):\n    return np.sqrt(mean_squared_error(y_true, y_pred))","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:24.88944Z","iopub.execute_input":"2025-01-02T06:01:24.889706Z","iopub.status.idle":"2025-01-02T06:01:24.894145Z","shell.execute_reply.started":"2025-01-02T06:01:24.889682Z","shell.execute_reply":"2025-01-02T06:01:24.893523Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# 검증 세트에 대한 예측\nval_pred = lr.predict(val_x)\n\n# RMSLE 점수 계산\nrmsle_score = rmsle(val_y, val_pred)\nprint(f'검증 세트 RMSLE: {rmsle_score:.4f}')","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:24.895728Z","iopub.execute_input":"2025-01-02T06:01:24.895967Z","iopub.status.idle":"2025-01-02T06:01:24.950787Z","shell.execute_reply.started":"2025-01-02T06:01:24.895946Z","shell.execute_reply":"2025-01-02T06:01:24.949495Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# 테스트 세트 예측 및 제출 파일 생성\ntest_pred = lr.predict(test_processed)\nsubmission['Premium Amount'] = np.expm1(test_pred)  # log 변환 되돌리기\nsubmission.to_csv('submission.csv', index=False)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:24.951758Z","iopub.execute_input":"2025-01-02T06:01:24.952162Z","iopub.status.idle":"2025-01-02T06:01:26.456869Z","shell.execute_reply.started":"2025-01-02T06:01:24.952123Z","shell.execute_reply":"2025-01-02T06:01:26.456138Z"}},"outputs":[],"execution_count":null},{"cell_type":"markdown","source":"![image.png](attachment:image.png)","metadata":{},"attachments":{"image.png":{"image/png":"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"}}},{"cell_type":"markdown","source":"### Modeling","metadata":{}},{"cell_type":"code","source":"train=pd.read_csv(data_path+'train.csv')\ntest=pd.read_csv(data_path+'test.csv')\nsubmission=pd.read_csv(data_path+'sample_submission.csv')\n\ntrain=train.drop('id',axis=1)\ntest=test.drop('id',axis=1)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:26.457584Z","iopub.execute_input":"2025-01-02T06:01:26.457785Z","iopub.status.idle":"2025-01-02T06:01:32.670083Z","shell.execute_reply.started":"2025-01-02T06:01:26.457767Z","shell.execute_reply":"2025-01-02T06:01:32.669434Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"all_data=pd.concat([train, test],ignore_index=True)\nall_data=all_data.drop('Premium Amount', axis=1)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:32.670845Z","iopub.execute_input":"2025-01-02T06:01:32.67113Z","iopub.status.idle":"2025-01-02T06:01:33.317025Z","shell.execute_reply.started":"2025-01-02T06:01:32.671103Z","shell.execute_reply":"2025-01-02T06:01:33.316328Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# 결측치 확인\nprint(\"처리 전 결측치:\", all_data.isnull().sum().sum())\n\n# 데이터 타입별로 결측치 처리\nnumeric_columns = all_data.select_dtypes(include=['int64', 'float64']).columns\nobject_columns = all_data.select_dtypes(include=['object']).columns\n\n# 결측치 처리\nall_data[numeric_columns] = all_data[numeric_columns].fillna(all_data[numeric_columns].mean())\nall_data[object_columns] = all_data[object_columns].fillna('None')\n\n# 결측치 처리 확인\nprint(\"처리 후 결측치:\", all_data.isnull().sum().sum())\n\n# 각 컬럼별 결측치 확인\nprint(\"\\n각 컬럼별 결측치:\")\nprint(all_data.isnull().sum())","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:33.318002Z","iopub.execute_input":"2025-01-02T06:01:33.31831Z","iopub.status.idle":"2025-01-02T06:01:39.344261Z","shell.execute_reply.started":"2025-01-02T06:01:33.31828Z","shell.execute_reply":"2025-01-02T06:01:39.343334Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# # 결측치 처리\n# print(\"처리 전 결측치:\", all_data.isnull().sum().sum())\n\n# # 수치형 데이터 결측치 처리\n# all_data['Age'].fillna(all_data['Age'].mean(), inplace=True)\n# all_data['Annual Income'].fillna(all_data.groupby('Education Level')['Annual Income'].transform('mean'), inplace=True)\n# all_data['Number of Dependents'].fillna(0, inplace=True)\n# all_data['Health Score'].fillna(all_data['Health Score'].mean(), inplace=True)\n# all_data['Previous Claims'].fillna(all_data['Previous Claims'].mode()[0], inplace=True)\n# all_data['Vehicle Age'].fillna(0, inplace=True)\n# all_data['Credit Score'].fillna(all_data['Credit Score'].mean(), inplace=True)\n# all_data['Insurance Duration'].fillna(all_data['Insurance Duration'].mode()[0], inplace=True)\n\n# # 범주형 데이터 결측치 처리\n# all_data['Marital Status'].fillna('Single', inplace=True)\n# all_data['Occupation'].fillna('None', inplace=True)\n# all_data['Customer Feedback'].fillna('Average', inplace=True)\n\n# print(\"처리 후 결측치:\", all_data.isnull().sum().sum())\n\n# # 각 컬럼별 결측치 확인\n# print(\"\\n각 컬럼별 결측치:\")\n# print(all_data.isnull().sum())","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:39.345219Z","iopub.execute_input":"2025-01-02T06:01:39.345562Z","iopub.status.idle":"2025-01-02T06:01:39.349635Z","shell.execute_reply.started":"2025-01-02T06:01:39.345531Z","shell.execute_reply":"2025-01-02T06:01:39.348831Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# numeric_columns=all_data.select_dtypes(include=['int64', 'float64']).columns\n# object_columns=all_data.select_dtypes(include=['object']).columns","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:39.350479Z","iopub.execute_input":"2025-01-02T06:01:39.350782Z","iopub.status.idle":"2025-01-02T06:01:39.363648Z","shell.execute_reply.started":"2025-01-02T06:01:39.350753Z","shell.execute_reply":"2025-01-02T06:01:39.362862Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# all_data[numeric_columns].fillna(0, inplace=True)\n# all_data[object_columns].fillna('None', inplace=True)\n\n# print(all_data.isnull().sum().sum())\n\n# resumetable(all_data)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:39.364453Z","iopub.execute_input":"2025-01-02T06:01:39.364752Z","iopub.status.idle":"2025-01-02T06:01:39.377707Z","shell.execute_reply.started":"2025-01-02T06:01:39.364724Z","shell.execute_reply":"2025-01-02T06:01:39.37697Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"def encode_binary(df):\n    gender_map={'Male':0, 'Female':1}\n    smoking_map={'No':0, 'Yes':1}\n    \n    df['Gender']=df['Gender'].map(gender_map)\n    df['Smoking Status']=df['Smoking Status'].map(smoking_map)\n    \n    return df","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:39.378526Z","iopub.execute_input":"2025-01-02T06:01:39.378816Z","iopub.status.idle":"2025-01-02T06:01:39.394349Z","shell.execute_reply.started":"2025-01-02T06:01:39.378787Z","shell.execute_reply":"2025-01-02T06:01:39.393604Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"def encode_ordinal(df):\n    education_map={'High School':0, \"Bachelor's\":1, \"Master's\":2, \"PhD\":3}\n    policy_map={'Basic':0, 'Comprehensive':1, 'Premium':2}\n    feedback_map={'Poor':0, 'Average':1, 'Good':2}\n    exercise_map={'Rarely':0, 'Monthly':1, 'Weekly':2, 'Daily':3}\n    \n    df['Education Level']=df['Education Level'].map(education_map)\n    df['Policy Type']=df['Policy Type'].map(policy_map)\n    df['Customer Feedback']=df['Customer Feedback'].map(feedback_map)\n    df['Exercise Frequency']=df['Exercise Frequency'].map(exercise_map)\n    \n    return df","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:39.395204Z","iopub.execute_input":"2025-01-02T06:01:39.395458Z","iopub.status.idle":"2025-01-02T06:01:39.407949Z","shell.execute_reply.started":"2025-01-02T06:01:39.395428Z","shell.execute_reply":"2025-01-02T06:01:39.407195Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"all_data=encode_binary(all_data)\nall_data=encode_ordinal(all_data)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:39.408696Z","iopub.execute_input":"2025-01-02T06:01:39.408983Z","iopub.status.idle":"2025-01-02T06:01:40.066111Z","shell.execute_reply.started":"2025-01-02T06:01:39.408955Z","shell.execute_reply":"2025-01-02T06:01:40.065425Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"datetime_col=pd.to_datetime(all_data['Policy Start Date'])\n\nall_data.insert(15, 'Year', datetime_col.dt.year)\nall_data.insert(16, 'Month', datetime_col.dt.month)\nall_data.insert(17, 'Day', datetime_col.dt.day)\n\nall_data=all_data.drop('Policy Start Date', axis=1)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:40.066899Z","iopub.execute_input":"2025-01-02T06:01:40.067113Z","iopub.status.idle":"2025-01-02T06:01:41.134474Z","shell.execute_reply.started":"2025-01-02T06:01:40.067095Z","shell.execute_reply":"2025-01-02T06:01:41.133744Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"all_data.isnull().sum()","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:41.135189Z","iopub.execute_input":"2025-01-02T06:01:41.135413Z","iopub.status.idle":"2025-01-02T06:01:41.502179Z","shell.execute_reply.started":"2025-01-02T06:01:41.135395Z","shell.execute_reply":"2025-01-02T06:01:41.501296Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"from sklearn.preprocessing import OneHotEncoder\nonehot_list=['Marital Status','Occupation','Location','Property Type']\n\noh=OneHotEncoder()\nonehot_cat=oh.fit_transform(all_data[onehot_list])\nall_data=all_data.drop(onehot_list,axis=1)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:41.503164Z","iopub.execute_input":"2025-01-02T06:01:41.503515Z","iopub.status.idle":"2025-01-02T06:01:43.233631Z","shell.execute_reply.started":"2025-01-02T06:01:41.503482Z","shell.execute_reply":"2025-01-02T06:01:43.232908Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"all_data.shape, onehot_cat.shape","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:43.234339Z","iopub.execute_input":"2025-01-02T06:01:43.234561Z","iopub.status.idle":"2025-01-02T06:01:43.239386Z","shell.execute_reply.started":"2025-01-02T06:01:43.234544Z","shell.execute_reply":"2025-01-02T06:01:43.238776Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"onehot_cat","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:43.240208Z","iopub.execute_input":"2025-01-02T06:01:43.240465Z","iopub.status.idle":"2025-01-02T06:01:43.254914Z","shell.execute_reply.started":"2025-01-02T06:01:43.240446Z","shell.execute_reply":"2025-01-02T06:01:43.25413Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"from scipy import sparse\nall_data_sparse=sparse.hstack([sparse.csr_matrix(all_data), onehot_cat],format='csr')\nall_data_sparse.shape","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:43.255618Z","iopub.execute_input":"2025-01-02T06:01:43.255861Z","iopub.status.idle":"2025-01-02T06:01:44.873802Z","shell.execute_reply.started":"2025-01-02T06:01:43.255842Z","shell.execute_reply":"2025-01-02T06:01:44.873039Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"import numpy as np\nfrom sklearn.model_selection import train_test_split\n\nX_train=all_data_sparse[:len(train)]\nX_test=all_data_sparse[len(train):]\n\nprint(X_train.shape, X_test.shape)\n\ny_train=np.log1p(train['Premium Amount'])","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:44.874522Z","iopub.execute_input":"2025-01-02T06:01:44.874727Z","iopub.status.idle":"2025-01-02T06:01:45.395675Z","shell.execute_reply.started":"2025-01-02T06:01:44.874708Z","shell.execute_reply":"2025-01-02T06:01:45.394753Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"def rmsle(y_true, y_pred):\n    return np.sqrt(mean_squared_error(y_true, y_pred))","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:45.396443Z","iopub.execute_input":"2025-01-02T06:01:45.396658Z","iopub.status.idle":"2025-01-02T06:01:45.400387Z","shell.execute_reply.started":"2025-01-02T06:01:45.39664Z","shell.execute_reply":"2025-01-02T06:01:45.399462Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# from sklearn.model_selection import KFold, GridSearchCV\n# from catboost import CatBoostRegressor\n# from sklearn.metrics import make_scorer\n\n# # RMSLE scorer 생성\n# rmsle_scorer = make_scorer(rmsle, greater_is_better=False)\n\n# # K-fold 설정\n# n_splits = 5\n# kf = KFold(n_splits=n_splits, shuffle=True, random_state=42)\n\n# # 하이퍼파라미터 그리드 정의\n# param_grid = {\n#     'iterations': [1000],\n#     'learning_rate': [0.01, 0.1],\n#     'depth': [6, 8],\n#     'l2_leaf_reg': [1, 3],\n#     'bootstrap_type': ['Bernoulli'],\n#     'subsample': [0.8]\n# }\n\n# # CatBoostRegressor 및 GridSearchCV 설정\n# cat = CatBoostRegressor(\n#     random_seed=42,\n#     verbose=100\n# )\n\n# grid_search = GridSearchCV(\n#     estimator=cat,\n#     param_grid=param_grid,\n#     cv=kf,\n#     scoring=rmsle_scorer,  # RMSLE scorer 사용\n#     n_jobs=-1,\n#     verbose=2\n# )\n\n# # 학습 진행\n# print(\"그리드 서치 시작...\")\n# grid_search.fit(X_train, y_train)\n\n# print(\"\\n최적 파라미터:\", grid_search.best_params_)\n# print(\"최적 RMSLE 점수: {:.4f}\".format(-grid_search.best_score_))  # 음수 값을 양수로 변환","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:45.401127Z","iopub.execute_input":"2025-01-02T06:01:45.401334Z","iopub.status.idle":"2025-01-02T06:01:45.417436Z","shell.execute_reply.started":"2025-01-02T06:01:45.401317Z","shell.execute_reply":"2025-01-02T06:01:45.416686Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# from bayes_opt import BayesianOptimization\n# from catboost import CatBoostRegressor\n# from sklearn.model_selection import KFold\n# import numpy as np\n\n# # 평가 함수 정의\n# def cat_cv(learning_rate, depth, l2_leaf_reg, subsample, min_data_in_leaf, max_bin):\n#     # 파라미터 설정\n#     params = {\n#         'iterations': 2000,  # 증가\n#         'learning_rate': learning_rate,\n#         'depth': int(depth),\n#         'l2_leaf_reg': l2_leaf_reg,\n#         'bootstrap_type': 'Bernoulli',\n#         'subsample': subsample,\n#         'min_data_in_leaf': int(min_data_in_leaf),\n#         'max_bin': int(max_bin),\n#         'random_seed': 42,\n#         'verbose': 0,\n#         'task_type': 'GPU',\n#         'devices': '0',\n#         'gpu_ram_part': 0.95\n#     }\n    \n#     # K-fold 교차 검증\n#     n_splits = 5\n#     kf = KFold(n_splits=n_splits, shuffle=True, random_state=42)\n#     scores = []\n    \n#     for train_idx, val_idx in kf.split(X_train):\n#         train_data = X_train[train_idx]\n#         val_data = X_train[val_idx]\n#         train_target = y_train[train_idx]\n#         val_target = y_train[val_idx]\n        \n#         model = CatBoostRegressor(**params)\n#         model.fit(\n#             train_data, \n#             train_target,\n#             eval_set=[(val_data, val_target)],\n#             early_stopping_rounds=100,  # 증가\n#             verbose=False\n#         )\n        \n#         val_pred = model.predict(val_data)\n#         score = -rmsle(val_target, val_pred)\n#         scores.append(score)\n    \n#     return np.mean(scores)\n\n# # 베이지안 최적화 파라미터 범위 설정 (확장)\n# pbounds = {\n#     'learning_rate': (0.001, 0.3),\n#     'depth': (4, 12),\n#     'l2_leaf_reg': (0.01, 10.0),\n#     'subsample': (0.5, 1.0),\n#     'min_data_in_leaf': (1, 100),\n#     'max_bin': (200, 500)\n# }\n\n# # 베이지안 최적화 실행\n# optimizer = BayesianOptimization(\n#     f=cat_cv,\n#     pbounds=pbounds,\n#     random_state=42,\n#     verbose=50  # 더 자세한 로그 출력\n# )\n\n# # 탐색 전략 설정\n# from bayes_opt.logger import JSONLogger\n# from bayes_opt.event import Events\n# from bayes_opt.util import load_logs\n\n# # JSON 로거 설정\n# logger = JSONLogger(path=\"./catboost_optimization.json\")\n# optimizer.subscribe(Events.OPTIMIZATION_STEP, logger)\n\n# # 최적화 진행 (반복 횟수 증가)\n# print(\"베이지안 최적화 시작...\")\n# optimizer.maximize(\n#     init_points=10,    # 초기 랜덤 탐색 횟수 증가\n#     n_iter=50,        # 베이지안 최적화 반복 횟수 증가\n#     acq='ei',         # acquisition function: 'ei', 'ucb', 'poi'\n#     xi=0.01           # exploration-exploitation 트레이드오프 파라미터\n# )\n\n# # 최적 파라미터 및 중간 결과 출력\n# print(\"\\n=== 최적화 결과 ===\")\n# print(\"\\n최적 파라미터:\")\n# print(optimizer.max['params'])\n# print(f\"최적 RMSLE 점수: {-optimizer.max['target']:.4f}\")\n\n# # 상위 5개 결과 출력\n# print(\"\\n상위 5개 결과:\")\n# for i, res in enumerate(sorted(optimizer.res, key=lambda x: x['target'], reverse=True)[:5]):\n#     print(f\"\\n{i+1}번째 최적 결과:\")\n#     print(f\"RMSLE: {-res['target']:.4f}\")\n#     print(\"파라미터:\")\n#     for param, val in res['params'].items():\n#         print(f\"{param}: {val}\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:45.418094Z","iopub.execute_input":"2025-01-02T06:01:45.418353Z","iopub.status.idle":"2025-01-02T06:01:45.433316Z","shell.execute_reply.started":"2025-01-02T06:01:45.418322Z","shell.execute_reply":"2025-01-02T06:01:45.432668Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# # 최종 모델 학습\n# from catboost import CatBoostRegressor\n\n# # 최적 파라미터로 모델 설정\n# final_params = {\n#     'iterations': 2000,\n#     'learning_rate': optimizer.max['params']['learning_rate'],\n#     'depth': int(optimizer.max['params']['depth']),\n#     'l2_leaf_reg': optimizer.max['params']['l2_leaf_reg'],\n#     'bootstrap_type': 'Bernoulli',\n#     'subsample': optimizer.max['params']['subsample'],\n#     'min_data_in_leaf': int(optimizer.max['params']['min_data_in_leaf']),\n#     'max_bin': int(optimizer.max['params']['max_bin']),\n#     'random_seed': 42,\n#     'verbose': 100,\n#     'task_type': 'GPU',\n#     'devices': '0',\n#     'gpu_ram_part': 0.95\n# }\n\n# # 최종 모델 학습\n# print(\"최종 모델 학습 시작...\")\n# final_model = CatBoostRegressor(**final_params)\n# final_model.fit(\n#     X_train, \n#     y_train,\n#     eval_set=[(X_train, y_train)],\n#     early_stopping_rounds=100,\n#     verbose=100\n# )\n\n# # 테스트 데이터 예측\n# print(\"\\n테스트 데이터 예측 중...\")\n# test_pred = final_model.predict(X_test)\n\n# # 예측값 변환 (log 변환 되돌리기)\n# submission['Premium Amount'] = np.expm1(test_pred)\n\n# # 제출 파일 저장\n# submission.to_csv('submission.csv', index=False)\n# print(\"\\n제출 파일 생성 완료!\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:45.434042Z","iopub.execute_input":"2025-01-02T06:01:45.434224Z","iopub.status.idle":"2025-01-02T06:01:45.45025Z","shell.execute_reply.started":"2025-01-02T06:01:45.434208Z","shell.execute_reply":"2025-01-02T06:01:45.44955Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"import lightgbm as lgb\nimport xgboost as xgb\nfrom catboost import CatBoostRegressor\nimport numpy as np\nfrom sklearn.model_selection import KFold\nfrom bayes_opt import BayesianOptimization\n\n# LGBM 하이퍼파라미터 최적화 함수\ndef lgb_cv(learning_rate, num_leaves, max_depth, min_child_samples, subsample, colsample_bytree):\n    params = {\n        'objective': 'regression',\n        'metric': 'rmse',\n        'learning_rate': learning_rate,\n        'num_leaves': int(num_leaves),\n        'max_depth': int(max_depth),\n        'min_child_samples': int(min_child_samples),\n        'subsample': subsample,\n        'colsample_bytree': colsample_bytree,\n        'min_split_gain': 0.01,  # 추가: 최소 분할 이득 설정\n        'verbosity': -1,  # 추가: 경고 메시지 억제\n        'device': 'gpu',  # GPU 사용 설정\n        'gpu_platform_id': 0,  # GPU 플랫폼 ID (필요에 따라 조정)\n        'gpu_device_id': 0  # GPU 장치 ID (필요에 따라 조정)\n    }\n    \n    scores = []\n    kf = KFold(n_splits=5, shuffle=True, random_state=42)\n    \n    for train_idx, val_idx in kf.split(X_train):\n        train_data = lgb.Dataset(X_train[train_idx], y_train[train_idx])\n        val_data = lgb.Dataset(X_train[val_idx], y_train[val_idx])\n        \n        callbacks = [lgb.early_stopping(stopping_rounds=50)]\n        \n        model = lgb.train(\n            params,\n            train_data,\n            num_boost_round=2000,\n            valid_sets=[val_data],\n            callbacks=callbacks\n        )\n        \n        val_pred = model.predict(X_train[val_idx])\n        score = -rmsle(y_train[val_idx], val_pred)\n        scores.append(score)\n    \n    return np.mean(scores)\n\n# XGBoost 하이퍼파라미터 최적화 함수\ndef xgb_cv(learning_rate, max_depth, min_child_weight, subsample, colsample_bytree, gamma):\n    params = {\n        'objective': 'reg:squarederror',\n        'eval_metric': 'rmse',\n        'learning_rate': learning_rate,\n        'max_depth': int(max_depth),\n        'min_child_weight': min_child_weight,\n        'subsample': subsample,\n        'colsample_bytree': colsample_bytree,\n        'gamma': gamma,\n        'tree_method': 'gpu_hist',  # GPU 사용 설정\n        'predictor': 'gpu_predictor'  # GPU 예측기 설정\n    }\n    \n    scores = []\n    kf = KFold(n_splits=5, shuffle=True, random_state=42)\n    \n    for train_idx, val_idx in kf.split(X_train):\n        dtrain = xgb.DMatrix(X_train[train_idx], y_train[train_idx])\n        dval = xgb.DMatrix(X_train[val_idx], y_train[val_idx])\n        \n        model = xgb.train(\n            params,\n            dtrain,\n            num_boost_round=2000,\n            evals=[(dval, 'eval')],\n            early_stopping_rounds=50,\n            verbose_eval=False\n        )\n        \n        val_pred = model.predict(dval)\n        score = -rmsle(y_train[val_idx], val_pred)\n        scores.append(score)\n    \n    return np.mean(scores)\n\ndef cat_cv(learning_rate, depth, l2_leaf_reg, subsample, min_data_in_leaf, max_bin):\n    # 파라미터 설정\n    params = {\n        'iterations': 2000,  # 증가\n        'learning_rate': learning_rate,\n        'depth': int(depth),\n        'l2_leaf_reg': l2_leaf_reg,\n        'bootstrap_type': 'Bernoulli',\n        'subsample': subsample,\n        'min_data_in_leaf': int(min_data_in_leaf),\n        'max_bin': int(max_bin),\n        'random_seed': 42,\n        'task_type': 'GPU',  # GPU 사용 설정\n        'devices': '0',  # 사용할 GPU 장치 ID\n        'verbose': False  # 출력 메시지 억제\n    }\n    \n    # K-fold 교차 검증\n    n_splits = 5\n    kf = KFold(n_splits=n_splits, shuffle=True, random_state=42)\n    scores = []\n    \n    for train_idx, val_idx in kf.split(X_train):\n        train_data = X_train[train_idx]\n        val_data = X_train[val_idx]\n        train_target = y_train[train_idx]\n        val_target = y_train[val_idx]\n        \n        model = CatBoostRegressor(**params)\n        model.fit(\n            train_data, \n            train_target,\n            eval_set=[(val_data, val_target)],\n            early_stopping_rounds=100,  # 증가\n            verbose=False\n        )\n        \n        val_pred = model.predict(val_data)\n        score = -rmsle(val_target, val_pred)\n        scores.append(score)\n    \n    return np.mean(scores)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:45.451051Z","iopub.execute_input":"2025-01-02T06:01:45.451352Z","iopub.status.idle":"2025-01-02T06:01:48.847384Z","shell.execute_reply.started":"2025-01-02T06:01:45.451324Z","shell.execute_reply":"2025-01-02T06:01:48.846722Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# LGBM\nlgb_pbounds = {\n    'learning_rate': (0.001, 0.1),\n    'num_leaves': (20, 100),\n    'max_depth': (3, 12),\n    'min_child_samples': (1, 50),\n    'subsample': (0.5, 1.0),\n    'colsample_bytree': (0.5, 1.0)\n}\n\nlgb_optimizer = BayesianOptimization(\n    f=lgb_cv,\n    pbounds=lgb_pbounds,\n    random_state=42\n)\n\nlgb_optimizer.maximize(init_points=10, n_iter=50)\n\n# LightGBM\nlgb_params = lgb_optimizer.max['params']\nlgb_params['num_leaves'] = int(lgb_params['num_leaves'])\nlgb_params['max_depth'] = int(lgb_params['max_depth'])\nlgb_params['min_child_samples'] = int(lgb_params['min_child_samples'])\n\nlgb_train = lgb.Dataset(X_train, y_train)\nlgb_model = lgb.train(lgb_params, lgb_train, num_boost_round=2000)\nlgb_pred = lgb_model.predict(X_test)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T06:01:48.848183Z","iopub.execute_input":"2025-01-02T06:01:48.848753Z","iopub.status.idle":"2025-01-02T07:47:40.197723Z","shell.execute_reply.started":"2025-01-02T06:01:48.848717Z","shell.execute_reply":"2025-01-02T07:47:40.196926Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"pd.DataFrame({'id': submission['id'], 'Premium Amount': np.expm1(lgb_pred)}).to_csv('lgb_pred.csv', index=False)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T07:47:40.199966Z","iopub.execute_input":"2025-01-02T07:47:40.200207Z","iopub.status.idle":"2025-01-02T07:47:41.510788Z","shell.execute_reply.started":"2025-01-02T07:47:40.200187Z","shell.execute_reply":"2025-01-02T07:47:41.509892Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# XGBoost\nxgb_pbounds = {\n    'learning_rate': (0.001, 0.1),\n    'max_depth': (3, 12),\n    'min_child_weight': (1, 10),\n    'subsample': (0.5, 1.0),\n    'colsample_bytree': (0.5, 1.0),\n    'gamma': (0, 5)\n}\n\nxgb_optimizer = BayesianOptimization(\n    f=xgb_cv,\n    pbounds=xgb_pbounds,\n    random_state=42\n)\n\nxgb_optimizer.maximize(init_points=10, n_iter=50)\n\n# XGBoost\nxgb_params = xgb_optimizer.max['params']\nxgb_params['max_depth'] = int(xgb_params['max_depth'])\n\ndtrain = xgb.DMatrix(X_train, y_train)  # dtrain 변수 이름 수정\ndtest = xgb.DMatrix(X_test)\nxgb_model = xgb.train(xgb_params, dtrain, num_boost_round=2000)\nxgb_pred = xgb_model.predict(dtest)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T07:47:41.511751Z","iopub.execute_input":"2025-01-02T07:47:41.512014Z","iopub.status.idle":"2025-01-02T08:29:34.39563Z","shell.execute_reply.started":"2025-01-02T07:47:41.511994Z","shell.execute_reply":"2025-01-02T08:29:34.394908Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"pd.DataFrame({'id': submission['id'], 'Premium Amount': np.expm1(xgb_pred)}).to_csv('xgb_pred.csv', index=False)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T08:29:34.396144Z","iopub.execute_input":"2025-01-02T08:29:34.396343Z","iopub.status.idle":"2025-01-02T08:29:35.345998Z","shell.execute_reply.started":"2025-01-02T08:29:34.396325Z","shell.execute_reply":"2025-01-02T08:29:35.345333Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"cat_pbounds = {\n    'learning_rate': (0.001, 0.3),\n    'depth': (4, 12),\n    'l2_leaf_reg': (0.01, 10.0),\n    'subsample': (0.5, 1.0),\n    'min_data_in_leaf': (1, 100),\n    'max_bin': (200, 500)\n}\n\n# 베이지안 최적화 실행\noptimizer = BayesianOptimization(\n    f=cat_cv,\n    pbounds=cat_pbounds,\n    random_state=42\n)\n\noptimizer.maximize(\n    init_points=10,    # 초기 랜덤 탐색 횟수 증가\n    n_iter=50        # 베이지안 최적화 반복 횟수 증가\n)\n\nfinal_params = {\n    'iterations': 2000,\n    'learning_rate': optimizer.max['params']['learning_rate'],\n    'depth': int(optimizer.max['params']['depth']),\n    'l2_leaf_reg': optimizer.max['params']['l2_leaf_reg'],\n    'bootstrap_type': 'Bernoulli',\n    'subsample': optimizer.max['params']['subsample'],\n    'min_data_in_leaf': int(optimizer.max['params']['min_data_in_leaf']),\n    'max_bin': int(optimizer.max['params']['max_bin']),\n    'random_seed': 42\n}\n\n# CatBoost\ncatboost_model = CatBoostRegressor(**final_params)\ncatboost_model.fit(X_train, y_train)\ncatboost_pred = catboost_model.predict(X_test)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T08:49:03.142406Z","iopub.execute_input":"2025-01-02T08:49:03.14275Z","iopub.status.idle":"2025-01-02T09:26:50.054675Z","shell.execute_reply.started":"2025-01-02T08:49:03.142726Z","shell.execute_reply":"2025-01-02T09:26:50.053908Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"pd.DataFrame({'id': submission['id'], 'Premium Amount': np.expm1(catboost_pred)}).to_csv('catboost_pred.csv', index=False)","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T09:26:50.055768Z","iopub.execute_input":"2025-01-02T09:26:50.05598Z","iopub.status.idle":"2025-01-02T09:26:51.360698Z","shell.execute_reply.started":"2025-01-02T09:26:50.055962Z","shell.execute_reply":"2025-01-02T09:26:51.359969Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"# 앙상블 (가중평균)\nweights = [0.4, 0.3, 0.3]  # CatBoost, LightGBM, XGBoost 순서\nensemble_pred = weights[0] * catboost_pred + weights[1] * lgb_pred + weights[2] * xgb_pred\nsubmission['Premium Amount'] = np.expm1(ensemble_pred)\nsubmission.to_csv('submission.csv', index=False)\n\nprint(\"모든 예측 파일이 저장되었습니다!\")\nprint(\"1. catboost_pred.csv\")\nprint(\"2. lgb_pred.csv\")\nprint(\"3. xgb_pred.csv\")\nprint(\"4. ensemble_submission.csv (가중평균 앙상블)\")","metadata":{"trusted":true,"execution":{"iopub.status.busy":"2025-01-02T09:26:51.362168Z","iopub.execute_input":"2025-01-02T09:26:51.362515Z","iopub.status.idle":"2025-01-02T09:26:52.753889Z","shell.execute_reply.started":"2025-01-02T09:26:51.362491Z","shell.execute_reply":"2025-01-02T09:26:52.753027Z"}},"outputs":[],"execution_count":null},{"cell_type":"code","source":"","metadata":{"trusted":true},"outputs":[],"execution_count":null}]}