machine-learning - 我在 RandomSearchCV 中不断收到 AttributeError

Question

x_tu = data_cls_tu.iloc[:,1:].values
y_tu = data_cls_tu.iloc[:,0].values

classifier = DecisionTreeClassifier()
parameters = [{"max_depth": [3,None],
               "min_samples_leaf": np.random.randint(1,9),
               "criterion": ["gini","entropy"]}]
randomcv = RandomizedSearchCV(estimator=classifier, param_distributions=parameters,
                              scoring='accuracy', cv=10, n_jobs=-1,
                              random_state=0)
randomcv.fit(x_tu, y_tu)



---------------------------------------------------------------------------
AttributeError                            Traceback (most recent call last)
<ipython-input-17-fa8376cb54b8> in <module>()
     11                               scoring='accuracy', cv=10, n_jobs=-1,
     12                               random_state=0)
---> 13 randomcv.fit(x_tu, y_tu)

~\Anaconda3\lib\site-packages\sklearn\model_selection\_search.py in fit(self, X, y, groups, **fit_params)
    616         n_splits = cv.get_n_splits(X, y, groups)
    617         # Regenerate parameter iterable for each fit
--> 618         candidate_params = list(self._get_param_iterator())
    619         n_candidates = len(candidate_params)
    620         if self.verbose > 0:

~\Anaconda3\lib\site-packages\sklearn\model_selection\_search.py in __iter__(self)
    236         # in this case we want to sample without replacement
    237         all_lists = np.all([not hasattr(v, "rvs")
--> 238                             for v in self.param_distributions.values()])
    239         rnd = check_random_state(self.random_state)
    240 

AttributeError: 'list' object has no attribute 'values'

嗨，我在 RandomSearchCV 的拟合方法上不断收到错误。

当我在 GridSearchCV 上使用它们时它有效，但 GridSearchCV 花了 5 个小时才完成。

x_tu、y_tu 都是 numpy.ndarray 类型。

Answer 1

param_distributions 必须是 dict 对象 ( documentation )，但您传递的是包含单个 dict 的列表。删除外部方括号然后它应该可以正常工作。

应该是这样的:

parameters = {"max_depth": [3,None],
               "min_samples_leaf": [np.random.randint(1,9)],
               "criterion": ["gini","entropy"]}