빅데이터분서기사 실기 - XGBoostRegressor

import pandas as pd
import numpy as np 

train = pd.read_csv("./house-prices-advanced-regression-techniques/train.csv")
test = pd.read_csv("./house-prices-advanced-regression-techniques/test.csv")
y_train = train['SalePrice']
# null data 확인
ex_cols = train.isnull().sum().sort_values(ascending=False).head(6).index.tolist()

# null data drop
train.drop(ex_cols,axis=1, inplace=True)
train.drop('Id',axis=1,inplace=True)
test.drop(ex_cols,axis=1, inplace=True)
test.drop('Id',axis=1,inplace=True)
train.drop("SalePrice", axis=1, inplace=True)

# feature
obj_cols = train.select_dtypes(include='object').columns.tolist()
num_cols = [x for x in train.columns.tolist() if x not in obj_cols]
for col in obj_cols:
    train[col].fillna(train[col].value_counts().index[0], inplace=True) # 최빈값
    test[col].fillna(train[col].value_counts().index[0], inplace=True) # 최빈값
for col in num_cols:
    train[col].fillna(train[col].mean(), inplace=True)
    test[col].fillna(train[col].mean(), inplace=True)
print(train.shape)
#print(train.isnull().sum().sort_values(ascending=False).head(20))

# 레이블인코딩, minmax scaler
from sklearn.preprocessing import LabelEncoder
from sklearn.preprocessing import MinMaxScaler
le = LabelEncoder()
mm = MinMaxScaler()
for col in obj_cols:
    le.fit(train[col])
    train[col] = le.transform(train[col])
    test[col] = le.transform(test[col])
for col in num_cols:
    mm.fit(np.array(train[col]).reshape(-1,1))
    train[col] = mm.transform(np.array(train[col]).reshape(-1,1))
    test[col] = mm.transform(np.array(test[col]).reshape(-1,1))
    
# train,valid set 분리
from sklearn.model_selection import train_test_split
X_train, X_valid, y_train, y_valid = train_test_split(train, y_train, test_size=0.3)
print(X_train.shape, X_valid.shape, y_train.shape, y_valid.shape)

# xgboost, 평가
import xgboost as xgb
from sklearn.metrics import r2_score
xgb_model = xgb.XGBRegressor(n_estimators=300, max_depth=3, learning_rate=0.05, n_jobs=-1)
xgb_model.fit(X_train, y_train, eval_set = [(X_valid, y_valid)], early_stopping_rounds=30, verbose=0)

# 결과해석
pred = xgb_model.predict(X_valid)
import matplotlib.pyplot as plt
fig, (ax1, ax2) = plt.subplots(1,2, figsize=(12,4))
ax1.scatter(y_valid, pred)
ax1.set_xlabel("y_valid")
ax1.set_ylabel("prediction")
ax2.scatter(range(y_valid.shape[0]), pred-y_valid, label='pred-y_valid')
ax2.legend()
ax1.set_title("XGBOOST REGRESSOR")