python实现Kmeans文本聚类，通过PCA降维和Matplotlib显示聚类3d三维图像

首先感谢Eastmount写的内容http://blog.csdn.net/Eastmount/article/details/50545937。点击打开链接

在此基础上，主要实现以下改进及结果

1.替换使用sklearn.feature_extraction.text.TfidfVectorizer,将corpus文本转换为tfidf值的svm向量。

2.通过PCA降维和Matplotlib显示聚类3d三维图像，更容易区分各类。

3.jieba提取30个关键词，作为特征值，而非全文档内容，进行SVM。

目前刚到找工作中，希望大家看到后，能够给些建议到[email protected]，感谢！

具体代码如下：

#  coding:utf-8
"""
Created on 2016-05-03 @author:[email protected]

"""

import re
import codecs
import jieba.analyse
import matplotlib.pyplot as plt
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.cluster import KMeans
from sklearn.decomposition import PCA
from mpl_toolkits.mplot3d import Axes3D

if __name__ == "__main__":

    corpus = []
    for line in open('01_All_BHSpider_Content_Result.txt','r').readlines():  # 读取1000行字符串的txt文本
        line = line.strip()
        # 运用正则表达式,过滤英文及标点符号,中文标点符号
        line = line.decode("utf8")
        line = re.sub("[\s+\.\!\/_,$%^*(+\"\']+|[+——！，。？、~@#￥%……&*（）]+|[a-zA-Z]+".decode("utf8"),"/".decode("utf8"),line)
        # 用jieba的textrank算法,每行字符串提取30个特征词
        line_feature = jieba.analyse.textrank(line , topK=30)
        line = "/".join(line_feature)
        corpus.append(line)

    # 参考官方文档运用sklearn.feature_extraction.text.TfidfVectorizer,将corpus文本转换为tfidf值的svm向量
    tfidfvec = TfidfVectorizer()
    cop_tfidf = tfidfvec.fit_transform(corpus)
    weight = cop_tfidf.toarray()

    # KMeans聚类,确定为4类
    clf = KMeans(n_clusters=4)  #景区 动物 人物 国家
    s = clf.fit(weight)
    # 输出lable文档
    result = codecs.open("lable.txt",'w','utf-8')
    i=1
    while i<=len(clf.labels_):
        result.write(str(i)+','+str(clf.labels_[i-1])+'\r\n')
        i=i+1
    result.close()

    # PCA降维,绘制三维图
    pca = PCA(n_components=3)
    newData = pca.fit_transform(weight)

    x1 =[]
    y1 =[]
    z1 =[]
    i=0
    while i<400:
        x1.append(newData[i][0])
        y1.append(newData[i][1])
        z1.append(newData[i][2])
        i +=1

    x2 =[]
    y2 =[]
    z2=[]
    i=400
    while i<600:
        x2.append(newData[i][0])
        y2.append(newData[i][1])
        z2.append(newData[i][2])
        i +=1

    x3 =[]
    y3 =[]
    z3=[]
    i=600
    while i<800:
        x3.append(newData[i][0])
        y3.append(newData[i][1])
        z3.append(newData[i][2])
        i +=1

    x4 =[]
    y4 =[]
    z4=[]
    i=800
    while i<1000:
        x4.append(newData[i][0])
        y4.append(newData[i][1])
        z4.append(newData[i][2])
        i +=1

    fig=plt.figure()
    ax=fig.add_subplot(111,projection='3d')
    ax.scatter(x1,y1,z1,c='b')
    ax.scatter(x2,y2,z2,c='r')
    ax.scatter(x3,y3,z3,c='k')
    ax.scatter(x4,y4,z4,c='g')
    plt.show()