使用GridSearchCV筛选最优回归模型与缩放器接入ML流水线是否合理？

数据集用于预测连续浮点型结果，无分类任务，包含200个样本与50个特征。

import pandas as pd
from sklearn.neighbors import KNeighborsRegressor
from sklearn.gaussian_process import GaussianProcessRegressor
from sklearn.linear_model import LinearRegression, Lasso
from sklearn.pipeline import Pipeline
from sklearn.preprocessing import StandardScaler, QuantileTransformer, MinMaxScaler,  RobustScaler
from sklearn.model_selection import GridSearchCV, train_test_split
from sklearn.metrics import mean_absolute_error, mean_squared_error

• 实例化测试流水线：

test_pipe = Pipeline([
    ('scaler', StandardScaler()), 
    ('model', KNeighborsRegressor())])

• 构建参数网格并执行GridSearchCV：

param_grid = {'scaler': [MinMaxScaler(), RobustScaler(), QuantileTransformer(), StandardScaler(), 'passthrough'],
              'model': [KNeighborsRegressor(), LinearRegression(), Lasso(alpha=0.1), Lasso(alpha=0.5), GaussianProcessRegressor()],
              }
grid = GridSearchCV(test_pipe, param_grid, cv=4)
grid.fit(X_train, y_train)
print(grid.best_params_)
print(grid.score(X_test, y_test))
grid.get_params()

输出[1]：

{'model': Lasso(alpha=0.1), 'scaler': QuantileTransformer()} 
0.9843408040290395

• 基于最优参数构建正式流水线：

pipe = Pipeline([
    ('scaler', grid.best_params_['scaler']), 
    ('model', grid.best_params_['model'])])

pipe.fit(X_train,y_train)

输出[2]：

Pipeline(steps=[('scaler', QuantileTransformer()), ('model', Lasso(alpha=0.1))])

这种方法可能存在问题，是否需要绘制各缩放器/模型组合的得分、MAE/MSE，并评估模型的欠拟合/过拟合情况？

print(f'Score: {pipe.score(X_test, y_test)}')
print(f'MAE: {mean_absolute_error(y_test, pipe.predict(X_test))}')
print(f'MSE: {mean_squared_error(y_test, pipe.predict(X_test))}')

输出[3]：

Score: 0.9843408040290395
MAE: 13.785801047280785
MSE: 352.94279168844093

内容的提问来源于stack exchange，提问作者Drew Sanislo