from sklearn.feature_extraction.text import CountVectorizer
from sklearn.feature_extraction.text import TfidfTransformer
from sklearn.feature_selection import SelectKBest, chi2
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.naive_bayes import MultinomialNB
import numpy as np
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn import metrics
from text.textpredict import *
from sklearn.cross_validation import * def chi22():
train_words=["急需 钱用 不用 出售 如图 价值 千多 便宜 出售 出售 急 ",
"读 读 重复 读好输 不变 绿 求高人 指点迷津 ",
"诚召搛只呆家小时工，全职妈妈、在校学生、在家待业者、上班族、游戏者皆可做！每天5",
"发福利了 火熱找小莳工，每天在綫2--3小莳，莳涧地點没限制，薪资鈤结80--150/",
"急招小时工，每天在綫2--3小拭，时间地点没限制，薪资日结80--150/天，适 急招小时工，每天在綫2--3小拭，时间地点没限制，薪资日结80--150/天，适合学生党，手机党，上班族，有空闲时间者，有興趣缪系，QQ（937117723）咨询，此处不回！！",
"发福利来 火熱找小莳工，每天在綫2--3小莳，莳涧地點没限制，薪资鈤结80--150/",
" 读 不好 呜呜 ","这句 话 总是 知道 连读 ","求 师傅 交 口语 求有 耐心 老师 基础 学 ",
"听到 读 "
]
train_tags=[1,0,1,1,1,1,0,0,0,0] """
##就提取了词频CountVectorizer
count_v1 = CountVectorizer(stop_words=None, max_df=0.5)
counts_train = count_v1.fit_transform(train_words)
##卡方检验chi,配合selectkbest 对特征进行选择
chi= SelectKBest(chi2,10)
mychi2 = chi.fit(counts_train, train_tags)
hi2_train = mychi2.transform(counts_train)
clf = MultinomialNB(alpha=0.01)
clf.fit(hi2_train, np.asarray(train_tags))
priediced = cross_val_predict(clf, hi2_train, train_tags)
print metrics.confusion_matrix(train_tags, priediced)
"""
##tf-idf
Tfidf = TfidfVectorizer()
tfidf_train = Tfidf.fit_transform(train_words)
clf = MultinomialNB(alpha=0.01)
clf.fit(tfidf_train, np.asarray(train_tags))
priediced = cross_val_predict(clf, tfidf_train, train_tags)
print metrics.confusion_matrix(train_tags, priediced) #print hi2_train chi22()