Grid search for SVC : IndexError: too many indices for array

Question

我正在尝试使用 GridSearchCV 为 SVC 寻找最佳参数。

from sklearn.svm import SVC
from sklearn import svm, grid_search
from sklearn.model_selection import GridSearchCV

param_grid = [
        {'C': [1,5,10,100]},
        ]
algo = SVC(kernel="poly",  degree=5, coef0=2)
grid_search = GridSearchCV(algo, param_grid, cv=3, scoring='neg_mean_squared_error')
grid_search.fit(X_train, y_train)
print(grid_search.best_params_) #line 162

我收到以下错误：

  File "main.py", line 162, in <module>
  IndexError: too many indices for array

当我不使用时 GridSearchCV 它有效：

from sklearn.svm import SVC
from sklearn import svm, grid_search
from sklearn.model_selection import GridSearchCV

algo = SVC(kernel="poly", C=1, degree=5, coef0=2)
algo.fit(X_train, y_train)
predict_test = algo.predict(X_test)
mse = mean_squared_error(y_test, predict_test)
rmse = np.sqrt(mse)
print(rmse)

我得到一个分数。

Answer 1

GridSearchCV.fit() 接受目标值作为形状 [n_samples] 或 [n_samples, n_output].

的 array-like y

在你的情况下，(892,)。因此，重塑 y_train:

y_train = y_train.reshape(892,)