python用kemeans对关键词进行分类

#计算关键词的tfidf
from sklearn.cluster import KMeans
import pymysql

def mysql():
    db = pymysql.connect(host='localhost', user='root', passwd='199126', db='kettle', port=3306, charset='utf8')
    cursor = db.cursor()
    sql1 = "SELECT guanjianci,id,jianjie FROM keji3_copy1 "

    cursor.execute(sql1)
    alldata = cursor.fetchall()
    corpus = []
    for s in alldata:
        guanjian = s[0]
        id = s[1]
        jianjie=s[2]

        corpus.append(jianjie)
    return corpus

def update(i,a):
    db = pymysql.connect(host='localhost', user='root', passwd='199126', db='kettle', port=3306, charset='utf8')
    cursor = db.cursor()
    try:
        sql1 = "UPDATE keji3 SET fenlei ='%s' where id = %s" % (i, a)

        cursor.execute(sql1)
        # 提交到数据库执行
        db.commit()
    except:
        print("2")

        # 打印结果
    print(i,a)
# !/usr/bin/env python
# -*- coding: utf-8 -*-
'''
Author: razrlele
Email: razrlele@gmail.com
'''
import jieba
from sklearn.feature_extraction.text import TfidfVectorizer
from sklearn.cluster import KMeans


def jieba_tokenize(text):
    return jieba.lcut(text)


tfidf_vectorizer = TfidfVectorizer(tokenizer=jieba_tokenize, lowercase=False)
'''
tokenizer: 指定分词函数
lowercase: 在分词之前将所有的文本转换成小写，因为涉及到中文文本处理，
所以最好是False
'''
text_list = mysql()
# 需要进行聚类的文本集
tfidf_matrix = tfidf_vectorizer.fit_transform(text_list)

num_clusters = 20
km_cluster = KMeans(n_clusters=num_clusters, max_iter=300, n_init=40, init='k-means++', n_jobs=-1)
'''
n_clusters: 指定K的值
max_iter: 对于单次初始值计算的最大迭代次数
n_init: 重新选择初始值的次数
init: 制定初始值选择的算法
n_jobs: 进程个数，为-1的时候是指默认跑满CPU
注意，这个对于单个初始值的计算始终只会使用单进程计算，
并行计算只是针对与不同初始值的计算。比如n_init=10，n_jobs=40, 
服务器上面有20个CPU可以开40个进程，最终只会开10个进程
'''
# 返回各自文本的所被分配到的类索引
result = km_cluster.fit_predict(tfidf_matrix)

a=0
for i in result:
    a=a+1
    update(i,a)