【Scikit-learn】糖尿病患者のデータセットで複数の機械学習モデルで特徴量の全ての組み合わせを検討［Python］

＜セル１＞

from sklearn.datasets import load_diabetes
import pandas as pd

diabetes = load_diabetes()

df = pd.DataFrame(diabetes.data, columns=diabetes.feature_names)

df["target"] = diabetes.target

df

実行結果

＜セル２＞

import itertools
from sklearn.model_selection import train_test_split
from sklearn.linear_model import LinearRegression
from sklearn.linear_model import Lasso
from sklearn.linear_model import ElasticNet
from sklearn.linear_model import Ridge
from sklearn.svm import SVR
from sklearn.metrics import r2_score
import numpy as np

models = ["LinearRegression", "Lasso", "ElasticNet", "RidgeRegression", "SVR"]

trial=10

score_list = []; combination_list = []; model_list = []

for i in range(1, len(diabetes.feature_names)+1):
    print(i)
    for comb in itertools.combinations(diabetes.feature_names, i):
#         print(comb)
        
        x = df.loc[:, comb]
        y = df.loc[:, "target"]
        
#         print(x)
        
        for mod in models:
            pred_score = []
        
            for t in range(trial):
        
                x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.2, train_size=0.8)
            
                if mod == "LinearRegression":
                    model = LinearRegression()
                elif mod == "Lasso":
                    model = Lasso()
                elif mod == "ElasticNet":
                    model = ElasticNet()
                elif mod == "RidgeRegression":
                    model = Ridge()
                elif mod == "SVR":
                    model = SVR()

                model.fit(x_train, y_train)
                pred = model.predict(x_test)
                pred_score.append(r2_score(y_test, pred))
#                 print(mod, r2_score(y_test, pred))

            pred_ave = np.average(np.array(pred_score))

    #         print(pred_ave)

            if len(score_list) < 100:
                score_list.append(pred_ave)
                combination_list.append(comb)
                model_list.append(mod)
            elif len(score_list) == 100:
                if np.min(score_list) < pred_ave:
                    del combination_list[np.argmin(score_list)]
                    del model_list[np.argmin(score_list)]
                    del score_list[np.argmin(score_list)]
                    combination_list.append(comb)
                    score_list.append(pred_ave)
                    model_list.append(mod)
                
results = pd.DataFrame()
results["Model_name"] = model_list
results["Combination"] = combination_list
results["Score"] = score_list
results = results.sort_values("Score", ascending=False)

results.head(10)

実行結果

＜セル３＞

x = df.loc[:, ["age", "sex", "bmi", "bp", "s1", "s2", "s3", "s4", "s5", "s6"]]
y = df.loc[:, "target"]

trial = 100

pred_score = []; pred_train_score = []

for i in range(trial):
    x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.2, train_size=0.8)
    
    model = LinearRegression()
    model.fit(x_train, y_train)
    pred = model.predict(x_test)
    pred_score.append(r2_score(y_test, pred))
    
pred_ave = np.average(np.array(pred_score))
pred_max = np.max(np.array(pred_score))
pred_min = np.min(np.array(pred_score))
pred_std = np.std(np.array(pred_score))
    
print(pred_ave, pred_max, pred_min, pred_std)

実行結果
0.4814766954911379 0.620899838132245 0.25614892681740276 0.0664189148607429

＜セル４＞

x = df.loc[:, ["age", "sex", "bmi", "bp", "s1", "s2", "s3", "s4", "s5", "s6"]]
y = df.loc[:, "target"]

trial = 10

pred_score = []; pred_train_score = []

for i in range(trial):
    x_train, x_test, y_train, y_test = train_test_split(x, y, test_size=0.2, train_size=0.8)
    
    model = Ridge()
    model.fit(x_train, y_train)
    pred = model.predict(x_test)
    pred_score.append(r2_score(y_test, pred))
    
pred_ave = np.average(np.array(pred_score))
pred_max = np.max(np.array(pred_score))
pred_min = np.min(np.array(pred_score))
pred_std = np.std(np.array(pred_score))
    
print(pred_ave, pred_max, pred_min, pred_std)

実行結果
0.4146294707871937 0.44320724340491413 0.3438771552368486 0.02616811676869301