데이터분석_공모전_DBI] 기법 정리 DecisionTreeRegressor & Random Search

기법 정리 DecisionTreeRegressor & Random Search

 

회귀 문제

LinearRegression

KNeighborsRegressor

DecisionTreeRegressor

RandomForestRegressor

XGBRegressor

 

DecisionTree, Random Search

 

1) 라이브러리 불러오기

import numpy as np
import pandas as pd
import matplotlib.pyplot as plt
import seaborn as sns

 

 

2) 데이터 준비

# x, y 분리

# medv : 본인 소유 주택 가
target = 'medv'

x = data.drop(target, axis=1)
y = data.loc[:, target]

# 학습용, 평가용 데이터 분리
from sklearn.model_selection import train_test_split

x_train, x_test, y_train, y_test = train_test_split(x, y, test_size = 0.3, random_state=1)

 

 

3) DecisionTreeRegressor

from sklearn.tree import DecisionTreeRegressor
from sklearn.model_selection import cross_val_score
from sklearn.metrics import mean_absolute_error, r2_score

model_dt = DecisionTreeRegressor(random_state=1)

# 성능예측

cv_score = cross_val_score(model_dt, x_train, y_train, cv=5)

# 결과 확인

print(cv_score)
print(cv_score.mean())

 

 

4) 모델 튜닝

from sklearn.tree import DecisionTree
from skelarn.model_selection import RandomizedSearchCV

# max depth : 1~50
params = {'max_depth' : range(1,50)}

model_dt = DecisionTreeRegressor(random_state)

model = RandomizeCV(nidek_dt,
                                param,
                                cv =5,
                                n_iter = 20,
                                scoring = 'r2')

# 결과 확인
** model.cv_results_['mean_test_score'] : 테스트로 얻은 성능
** model.best_parms_ : 최적의 파라미터
** model.best_score_ : 최고 성능
print(model.cv_results_['mean_test_score'])
print('최적 파라미터:', model.best_params_)
print('최고 성능:', model.best_score

 

# 변수 중요도

## model.best_estimator_
plt.barh(  y = list(x), width = model.best_estimator_.feature_importances_) # 최선의 파라미터로 된 DecisionTree이다.