ValueError：Scikit学习中的形状不一致错误Train_Test拆分

Question

我正在准备对csv文档进行一些预测，比较工作描述与工资结果。我已经将数据集分解为训练和测试，其中的特征是我正在处理的内容，目标是我所预测的。当我去打印并确认这些记录进行了适当分开后，我收到以下错误：ValueError异常：不一致的形状

我的代码和所产生的误差如下：

import csv 
import numpy as np 

# create posting & label list 
postList = [] 
labelList = [] 
filename = '\Users\yantezia.patrick\Downloads\Postings.csv' 
csvFile = csv.reader(open(filename, 'r'), delimiter=",") 
for row in csvFile: 
    postList.append(row[2]) 
    labelList.append(row[10]) #appending specific columns to specific list #these willbe labels 

# remove first row 
postList = postList[1:] #clearing out the header rows 
labelList = labelList[1:] 

temp = np.array([float(i) for i in labelList]) 
med = np.median(temp) 
for i, val in enumerate(labelList): 
    if float(val) >= med: 
     labelList[i] = 1 
    else: 
     labelList[i] = 0 

# subset list 
postList = postList[:100] 
labelList = labelList[:100] 
print postList[:2] 

from sklearn.feature_extraction.text import CountVectorizer, TfidfVectorizer 
import pandas as pd 

# create term matrix 
cv = CountVectorizer(lowercase=True, stop_words='english', ngram_range=(1,3), min_df=10) 
tfidf = TfidfVectorizer(lowercase=True, stop_words='english', ngram_range=(1,3), min_df=10) 

tf_dm = cv.fit_transform(postList) 
tfidf_dm = tfidf.fit_transform(postList) 

pd.DataFrame(tfidf_dm.toarray(),index=postList,columns=tfidf.get_feature_names()).head(10) 

tfidf.get_feature_names() 
tm = tm.toarray() 
print tf_dm 
tm = cv.fit(postList) 
print tm.vocabulary_ 
print tf_dm.shape 
print tfidf_dm.shape 

#add labels to word vector 
from sklearn.cross_validation import train_test_split 

features_train1 = train_test_split(tf_dm, labels, test_size=0.33, random_state=42) 
features_test1 = train_test_split(tf_dm, labels, test_size=0.33, random_state=42) 
target_train1 = train_test_split(tf_dm, labels, test_size=0.33, random_state=42) 
target_test1 = train_test_split(tf_dm, labels, test_size=0.33, random_state=42) 

features_train2 = train_test_split(tfidf_dm, labels, test_size=0.33, random_state=7) 
features_test2 = train_test_split(tfidf_dm, labels, test_size=0.33, random_state=7) 
target_train2 = train_test_split(tfidf_dm, labels, test_size=0.33, random_state=7) 
target_test2 = train_test_split(tfidf_dm, labels, test_size=0.33, random_state=7) 

print np.sum(target_train1) 
print np.sum(target_test1) 
print target_train1 
print target_test1 
--------------------------------------------------------------------------- 
ValueError        Traceback (most recent call last) 
<ipython-input-82-53ecd8559f48> in <module>() 
----> 1 print np.sum(target_train1) 
     2 print np.sum(target_test1) 
     3 print target_train1 
     4 print target_test1 

C:\Users\yantezia.patrick\AppData\Local\Continuum\Anaconda\lib\site-packages\numpy\core\fromnumeric.pyc in sum(a, axis, dtype, out, keepdims) 
    1707   except AttributeError: 
    1708    return _methods._sum(a, axis=axis, dtype=dtype, 
-> 1709         out=out, keepdims=keepdims) 
    1710   # NOTE: Dropping the keepdims parameters here... 
    1711   return sum(axis=axis, dtype=dtype, out=out) 

C:\Users\yantezia.patrick\AppData\Local\Continuum\Anaconda\lib\site-packages\numpy\core\_methods.pyc in _sum(a, axis, dtype, out, keepdims) 
    23 def _sum(a, axis=None, dtype=None, out=None, keepdims=False): 
    24  return um.add.reduce(a, axis=axis, dtype=dtype, 
---> 25        out=out, keepdims=keepdims) 
    26 
    27 def _prod(a, axis=None, dtype=None, out=None, keepdims=False): 

C:\Users\yantezia.patrick\AppData\Local\Continuum\Anaconda\lib\site-packages\scipy\sparse\compressed.pyc in __add__(self, other) 
    340   elif isspmatrix(other): 
    341    if (other.shape != self.shape): 
--> 342     raise ValueError("inconsistent shapes") 
    343 
    344    return self._binopt(other,'_plus_') 

ValueError: inconsistent shapes 

Answer 1

能否请你解释一下你想什么去做？ features_train1,features_test1,target_train1和target_test的定义是相同的。当你修改随机状态时，它们都将具有相同的内容。 你有没有试图看看这些是什么？ 请看the documentation of train_test_split。 features_train1是一个元组，其中前两个元素是tf_dm的分割，而后两个元素是labels的分割。

什么是你最有可能的意思做了

features_train1, features_test1, labels_train1, labels_test1 = train_test_split(tf_dm, labels, test_size=0.33, random_state=42)