如何从管道中使用的gridsearch中找到最佳参数？

Question

我正在尝试使用GridSearchCV找到最佳参数。

这是我的代码：

from sklearn.pipeline import Pipeline
from sklearn.model_selection import GridSearchCV
from sklearn.preprocessing import MaxAbsScaler
clf = SVC(kernel="rbf")
n_components = [25,30,35,40]
Cs = [ 0.1, 1, 10,15,20,25,30,50,60,70,100,150,200]
gammas = [0.001, 0.01,0.05,0.08, 0.1,0.3,0.5, 0.7,1]
scalar = MaxAbsScaler()
pipe = Pipeline(steps=[('transformer', scalar),('pca', pca), ('estimator', clf)])
estimator = GridSearchCV(pipe,
                         dict(pca__n_components=n_components,
                              estimator__C=Cs,estimator__gamma=gammas),cv=5,n_jobs=-1)

scores = cross_val_score(estimator, X, y, cv = 5)
print('average accuracy : ',np.array(scores).mean(),np.std(np.array(scores)))

我想知道通过网格搜索选择的参数的最佳组合。当我尝试estimator.best_params_时出现以下错误

AttributeError: 'GridSearchCV' object has no attribute 'best_params_'

Answer 1

尝试一下？

from sklearn.pipeline import Pipeline
from sklearn.model_selection import GridSearchCV
from sklearn.preprocessing import MaxAbsScaler

clf = SVC(kernel="rbf")
scalar = MaxAbsScaler()

paramgrid = {'n_components': [25,30,35,40], 
             'Cs': [ 0.1, 1, 10,15,20,25,30,50,60,70,100,150,200],
             'gammas': [0.001, 0.01,0.05,0.08, 0.1,0.3,0.5, 0.7,1],
            }

pipe = Pipeline(steps=[('transformer', scalar),
                       ('pca', pca), 
                       ('estimator', clf)])

estimator = GridSearchCV(pipe,
                         paramgrid,
                         cv=5,
                         n_jobs=-1)

scores = cross_val_score(estimator, X, y, cv = 5)
print('average accuracy : ',np.array(scores).mean(),np.std(np.array(scores)))