ValueError：尝试对测试集进行预测时尺寸不匹配答案

【问题标题】：ValueError: dimension mismatch when trying to make prediction on test setValueError：尝试对测试集进行预测时尺寸不匹配
【发布时间】：2019-05-08 15:51:59
【问题描述】：

我是机器学习的新手，正在努力让分类器使用测试数据集进行预测。

我认为错误维度不匹配是由于已将矢量化器与测试集相匹配，但我已经解决了这个问题，但我仍然遇到了问题。

该错误是由于矢量化器在某个地方被覆盖，我相信通过查看它，但我找不到在哪里...

非常感谢您的帮助，我已经在这方面工作了很长时间:)

import sqlalchemy
import pandas as pd
import numpy as np
from sklearn.model_selection import train_test_split
from sklearn.feature_extraction.text import CountVectorizer
from sklearn.feature_extraction.text import TfidfTransformer
from sklearn.feature_selection import SelectKBest
from sklearn.feature_selection import chi2
from sklearn.naive_bayes import MultinomialNB
from sklearn.metrics import accuracy_score
from sklearn import metrics
from sklearn.metrics import accuracy_score
from sklearn import metrics
import pickle

### Connect to MYSQL database
##
#
dbServerName = "localhost"
dbUser = "root"
dbPassword = "woodycool123"
dbName = "azure_support_tweets"

engine = sqlalchemy.create_engine('mysql+pymysql://root:woodycool123@localhost:3306/azure_support_tweets')
pd.set_option('display.max_colwidth', -1)
df = pd.read_sql_table("preprocessed_tweets", engine)
data = pd.DataFrame(df)

### Training and Test Data Split
##
#
features_train, features_test, labels_train, labels_test = train_test_split(data['text_tweet'], data['main_category'], random_state = 42, test_size=0.34)

### CountVectorizer
##
#
cv = CountVectorizer(ngram_range=(1,2), stop_words='english', min_df=3, max_df=0.50)
features_train_cv = cv.fit_transform(features_train)
# Uncomment to print a matrix count of tokens
# print(features_train_cv.toarray())
print("Feature Count\nCountVectorizer() #", len(cv.get_feature_names()))


### TF-IDF Transformer
##
#
tfidfv = TfidfTransformer(use_idf=True)
features_train_tfidfv = tfidfv.fit_transform(features_train_cv)
print("Feature Set\nTfidfVectorizer() #", features_train_tfidfv.shape)
# Remove to print the top 10 features
# features = tfidfv.get_feature_names()
# feature_order = np.argsort(tfidfv.idf_)[::-1]
# top_n = 10
# top_n_features = [features[i] for i in feature_order[:top_n]]
# print(top_n_features)


### SelectKBest
##
#
selector = SelectKBest(chi2, k=1000).fit_transform(features_train_tfidfv, labels_train)
print("Feature Set\nSelectKBest() and chi2 #", selector.shape)

### Train Model
##
#
clf = MultinomialNB()
clf.fit(selector, labels_train)


### Test Model
##
#
features_test_cv = cv.transform(features_test)
features_test_cv_two = tfidfv.transform(features_test_cv)
pred = clf.predict(features_test_cv)

错误：

Traceback (most recent call last):
  File "/Users/bethwalsh/Documents/classifier-twitter/building_the_classifer/feature_generation_selection.py", line 76, in <module>
    pred = clf.predict(features_test_cv)
  File "/Users/bethwalsh/anaconda3/lib/python3.6/site-packages/sklearn/naive_bayes.py", line 66, in predict
    jll = self._joint_log_likelihood(X)
  File "/Users/bethwalsh/anaconda3/lib/python3.6/site-packages/sklearn/naive_bayes.py", line 725, in _joint_log_likelihood
    return (safe_sparse_dot(X, self.feature_log_prob_.T) +
  File "/Users/bethwalsh/anaconda3/lib/python3.6/site-packages/sklearn/utils/extmath.py", line 135, in safe_sparse_dot
    ret = a * b
  File "/Users/bethwalsh/anaconda3/lib/python3.6/site-packages/scipy/sparse/base.py", line 515, in __mul__
    raise ValueError('dimension mismatch')
ValueError: dimension mismatch

【问题讨论】：

标签： python machine-learning scikit-learn classification

【解决方案1】：

看起来您忘记在测试模型部件中使用降维又名 SelectKBest。如果您要转换测试数据，我不知道以这种方式使用SelectKBest 是否正确。但无论如何，朴素贝叶斯模型

clf = MultinomialNB()
clf.fit(selector, labels_train)

等待selector 的形状，即在您的示例中 k=1000。在模型的测试部分，

features_test_cv = cv.transform(features_test)
features_test_cv_two = tfidfv.transform(features_test_cv)
pred = clf.predict(features_test_cv)

您跳过了这个转换，所以clf.predict 采用其他形状的矩阵。尝试使用SelectKBest.transform 来获得所需的输出：

selector_model = SelectKBest(chi2, k=1000). # create an object, use it later
selector = selector_model.fit_transform(features_train_tfidfv, labels_train)
print("Feature Set\nSelectKBest() and chi2 #", selector.shape)

clf = MultinomialNB()
clf.fit(selector, labels_train)

features_test_cv = cv.transform(features_test)
features_test_cv_two = tfidfv.transform(features_test_cv)
selector_test = selector_model.transform(features_test_cv_two)
pred = clf.predict(selector_test)

【讨论】：

【解决方案2】：

你也需要通过选择器通过测试集，但首先你必须适应

selector = SelectKBest(chi2, k=1000)
selector.fit(features_train_tfidfv, labels_train)

clf = MultinomialNB()
clf.fit(selector.transform(features_train_tfidfv), labels_train)

features_test_cv = selector.transform(tfidfv.transform(cv.transform(features_test)))
pred = clf.predict(features_test_cv)

之所以会抛出该错误，是因为选择器正在减少训练集的维度，而不是测试集的维度

【讨论】：