IndexError :索引越界

Question

我已经实现了MultinomialNB，但是我得到了这个消息。请帮我解决这个问题。下面是我的代码：

​

kf = KFold(len(X), n_folds=2, shuffle=True, random_state=9999)
model_train_index = []
model_test_index = []
model = 0

for k, (index_train, index_test) in enumerate(kf):
    X_train, X_test, y_train, y_test = X.ix[index_train,:], X.ix[index_test,:],y[index_train], y[index_test]
    clf = MultinomialNB(alpha=0.1).fit(X_train, y_train)
    score = clf.score(X_test, y_test)
    f1score = f1_score(y_test, clf.predict(X_test))
    precision = precision_score(y_test, clf.predict(X_test))
    recall = recall_score(y_test, clf.predict(X_test))
    print('Model %d has accuracy %f with | f1score: %f | precision: %f | recall : %f'%(k,score, f1score, precision, recall))
    model_train_index.append(index_train)
    model_test_index.append(index_test)
    model+=1

​

然后我得到这样的结果：

​

IndexError                                Traceback (most recent call last)
<ipython-input-3-df0b24edb687> in <module>()
      5 
      6 for k, (index_train, index_test) in enumerate(kf):
----> 7     X_train, X_test, y_train, y_test = X.ix[index_train,:], X.ix[index_test,:],y[index_train], y[index_test]
      8     clf = MultinomialNB(alpha=0.1).fit(X_train, y_train)
      9     score = clf.score(X_test, y_test)

IndexError: index 100 is out of bounds for axis 0 with size 100

​

Answer 1

Python使用从零开始的索引，因此如果X.ix[index_train,:]或y[index_train]的第零维是100，那么有效的index_train的最大值是99。index_test也是如此。

中的某些内容

kf = KFold(len(X), n_folds=2, shuffle=True, random_state=9999)

在枚举(Kf)时，会导致这些索引中的一个对于其中一个数组太大。