将GridSearchCV与sklearn.svm用于c和gamma时的值错误

时间:2016-02-23 12:19:53

标签: python scikit-learn svm cross-validation grid-search

clf = svm.SVC(kernel="rbf" , C = 1 , gamma = 1, probability = True)
from sklearn.grid_search import GridSearchCV

gamma_range = [1e-2,1e-1,1e0,1e1,1e2]
c_range = [1e-2,1e-1,1e0,1e1,1e2]
param_grid = dict(gamma = gamma_range, c = c_range)
print param_grid

给出:

{'c': [0.01, 0.1, 1.0, 10.0, 100.0], 'gamma': [0.01, 0.1, 1.0, 10.0, 100.0]}

正如所料,但是当我运行下一段代码时:

grid = GridSearchCV(clf, param_grid, cv= 10, scoring="accuracy")
grid.fit(X_norm, y)

我明白了:ValueError:估算器SVC的参数c无效

我非常确定这些是c的有效值,因为我已经通过自己手动运行它们,我只想微调一切。

0 个答案:

没有答案