python学习-文本数据分析2（文本分类）

利用Python进行文本分类, 
可用于过滤垃圾文本
1. 抽样
2. 人工标注样本文本中垃圾信息
3. 样本建模
4. 模型评估
5. 新文本预测
参考: 
http://scikit-learn.org/stable/user_guide.html
PYTHON自然语言处理中文翻译 NLTK Natural Language Processing with Python 中文版
主要步骤: 
1. 分词
2. 特征词提取
3. 生成词-文档矩阵
4. 整合分类变量
5. 建模
6. 评估

7. 预测新文本

#示例
#!/usr/bin/env python
# -*- coding:utf-8 -*-
import MySQLdb
import pandas as pd
import numpy as np
import jieba
import nltk
import jieba.posseg as pseg
from sklearn import cross_validation

#1. 读取数据,type为文本分类,0/1变量
df = pd.read_csv('F:\csv_test.csv',names=['id','cont','type'])

#2. 关键抽取
cont = df['cont']
tagall=[]
for t in cont:
        tags = jieba.analyse.extract_tags(t,kn)
        tagall.append(tags)
dist = nltk.FreqDist(tagall) #词频统计选top100的关键词
fea_words = fdist.keys()[:100]

#3. 生成词-文档矩阵
def word_features(content, top_words):
      word_set = set(content)
      features = {}
      for w in top_words:
          features["w_%s" % w] = (w in word_set)
      return features

#4. 整合矩阵与分类结果变量 
def data_feature(df, fea_words):
   data_set = []
   cont = df['cont']
   for i in range(0,len(cont)):
        content =jieba.cut(cont)
        feat = word_features(content,fea_words )
        category = df.loc[i,'type']
        tup = (feat, category)
        data_set.append(tup)
    return  data_set

data_list = data_feature(df, fea_words)
#5. 建立分类模型
#训练集与测试集
train_set,test_set = cross_validation.train_test_split(data_list,test_size=0.5)
#建模,贝叶斯
classifier = nltk.NaiveBayesClassifier.train(train_set)
#建模,决策树
classifier = nltk.DecisionTreeClassifier.train(train_set)

#6. 模型评估准确率
print nltk.classify.accuracy(classifier,test_set)

#7. 预测结果输出
pre_set = data_feature(new_data,fea_words)
pre_result = []
for item in pre_set:
    result = classifier.classify(item)
    pre_result.append(result)
#查看预测结果分布
pre_tab = set(pre_result)
for p in pre_tab:
    print p,pre_result.count(p)

其中2中特征词提取可采用各种方法进行, 
3,4步骤可改善,提高性能, 
5建模部分的模型可采用更多分类模型,逻辑回归,SVM...