索引错误:索引 12 超出大小为 12 的梯度提升分类器的轴 1 的界限

Question

>我试图在梯度提升分类上绘制错误，但我似乎找不到我的错误。我在网站上查找了类似的主题，但没有找到任何令人满意的答案。这是我的代码，希望你们能够提供帮助：

import time
tableau_duree_grd = np.zeros(145)
tableau_erreur_grd = np.zeros(145)
for b in range(5,150):
 start_time=time.time()
 grd=GradientBoostingClassifier(n_estimators=b,validation_fraction= 0.1,n_iter_no_change=10,learning_rate=0.1,max_features=None)
 grd.fit(XTrainD,YTrainD)
 pred = grd.predict(XTestD)
 test_erreur_grd = np.mean(YTestD!=pred)
 end_time=time.time()
 duree=end_time-start_time
 tableau_duree_grd[b-5]=duree
 tableau_erreur_grd[b-5]=test_erreur_grd

完整的错误回溯：

IndexError                                Traceback (most recent call last)
<ipython-input-46-9978ac0dd8ba> in <module>
  6     start_time=time.time()
  7     grd=GradientBoostingClassifier(n_estimators=b,validation_fraction= 0.1, n_iter_no_change=10,learning_rate=0.1,max_features=None)
----> 8     grd.fit(XTrainD,YTrainD)
  9     pred = grd.predict(XTestD)
 10     test_erreur_grd = np.mean(YTestD!=pred)
~AppDataLocalContinuumanaconda3libsite-packagessklearnensemblegradient_boosting.py in fit(self, X, y, sample_weight, monitor)
  1463         n_stages = self._fit_stages(X, y, y_pred, sample_weight, self._rng,
  1464                                     X_val, y_val, sample_weight_val,
-> 1465                                     begin_at_stage, monitor, X_idx_sorted)
  1466 
  1467         # change shape of arrays after fit (early-stopping or additional ests)
~AppDataLocalContinuumanaconda3libsite-packagessklearnensemblegradient_boosting.py in _fit_stages(self, X, y, y_pred, sample_weight, random_state, X_val, y_val, sample_weight_val, begin_at_stage, monitor, X_idx_sorted)
  1527             y_pred = self._fit_stage(i, X, y, y_pred, sample_weight,
  1528                                      sample_mask, random_state, X_idx_sorted,
-> 1529                                      X_csc, X_csr)
  1530 
  1531             # track deviance (= loss)
~AppDataLocalContinuumanaconda3libsite-packagessklearnensemblegradient_boosting.py in _fit_stage(self, i, X, y, y_pred, sample_weight, sample_mask, random_state, X_idx_sorted, X_csc, X_csr)
  1169 
  1170             residual = loss.negative_gradient(y, y_pred, k=k,
-> 1171                                               sample_weight=sample_weight)
  1172 
  1173             # induce regression tree on residuals
~AppDataLocalContinuumanaconda3libsite-packagessklearnensemblegradient_boosting.py in negative_gradient(self, y, pred, k, **kwargs)
   914             The index of the class
   915         """
--> 916         return y - np.nan_to_num(np.exp(pred[:, k] -
   917                                         logsumexp(pred, axis=1)))
   918 
  IndexError: index 12 is out of bounds for axis 1 with size 12

Answer 1

我不知道该软件，但这听起来像是一个错误。编程语言通常从 0 开始计算索引，而不是从 1 开始计算索引。

更具体地说，在gradient_boosting.py的第 916 行，"k"变量可能包含值 12，但应该使用值 11。

还可以通过在gradient_boosting.py的第 916 行周围添加 print 语句来获取更多信息，以便更好地了解发生错误时逻辑中发生的情况。