python-3.x - 如何将包含稀疏矩阵的 nd 数组拟合到 sklearn

Question

首先，我在我的数据框中的文本列上应用 TfidfVectorization 之后我使用 hsatck 合并它们

train_gene_var_tfidf = 
hstack((train_gene_feature_tfidf,train_variation_feature_tfidf))
test_gene_var_tfidf = 
hstack((test_gene_feature_tfidf,test_variation_feature_tfidf))
cv_gene_var_tfidf = 
hstack((cv_gene_feature_tfidf,cv_variation_feature_tfidf))

# Adding the train_text feature
train_x_tfidf = hstack((train_gene_var_tfidf, train_text))
train_x_tfidf = hstack((train_x_tfidf, train_text_feature_tfidf))
train_y = np.array(list(train_df['Class']))

结果我得到了一个包含稀疏矩阵的 numpy ndarray 我想将我的函数放入 sklearn .fit() 中，如下所示

clf = SGDClassifier(class_weight='balanced',alpha=i,penalty='l2', loss='log', random_state=42)
clf.fit(train_x_tfidf, train_y)

形状如下：

print(train_x_tfidf.shape)
print(train_y.shape)
>>>(4,)
>>>(2124,)
print(train_x_tfidf)
>>>[<2124x227 sparse matrix of type '<class 'numpy.float64'>'
with 2124 stored elements in Compressed Sparse Row format>
<2124x1000 sparse matrix of type '<class 'numpy.float64'>'
with 1259 stored elements in Compressed Sparse Row format>
<2124x1000 sparse matrix of type '<class 'numpy.float64'>'
with 31088 stored elements in Compressed Sparse Row format>
<2124x1000 sparse matrix of type '<class 'numpy.float64'>'
with 1129081 stored elements in Compressed Sparse Column format>]

我在尝试适应时遇到的错误是

ValueError：使用序列设置数组元素。

如何将此 ndarray 拟合到 sklearn 拟合函数中？

标签： python-3.xnumpymachine-learningscikit-learn