网格搜素

代码：

# -*- coding: utf-8 -*-
# 任务：鸢尾花识别
from sklearn.model_selection import train_test_split, GridSearchCV
from sklearn.neighbors import KNeighborsClassifier
from sklearn.linear_model import LogisticRegression
from sklearn.svm import SVC
from sklearn.datasets import load_iris
import warnings
warnings.filterwarnings(action="ignore")
def main():
    X,y = load_iris(return_X_y=True)
    X_train, X_test, y_train, y_test = train_test_split(X, y, test_size=1/3, random_state=10)
    model_dict = {'kNN':
                      (
                          KNeighborsClassifier(),
                          {'n_neighbors': [5, 15, 25], 'p': [1, 2]}
                       ),
                  'Logistic Regression':
                      (
                          LogisticRegression(),
                          {'C': [1e-2, 1, 1e2]}
                      ),
                  'SVM':
                      (
                          SVC(),
                          {'C': [1e-2, 1, 1e2]}
                      )
                  }   # 名称+元组
    for model_name, (model, model_params) in model_dict.items():
        # 训练模型
        clf = GridSearchCV(estimator=model, param_grid=model_params, cv=5) #模型、参数、折数
        clf.fit(X_train, y_train)   #训练
        best_model = clf.best_estimator_   #最佳模型的对象
        # 验证
        acc = best_model.score(X_test, y_test)
        print('{}模型的预测准确率：{:.2f}%'.format(model_name, acc * 100))
        print('{}模型的最优参数：{}'.format(model_name, clf.best_params_))       #最好的模型名称和参数
if __name__ == '__main__':
    main()

结果：