weka发现dblp论文合著者关系-数据挖掘作业

dblp数据集下载地址：https://dblp.org/xml/

拿到数据集首先要把他变成csv格式的文件

xml转json格式：

import xml.sax
from xml.sax.handler import ContentHandler
from xml.sax import parse
import json

doc = open('out2.json', 'w' , encoding='utf-8')
json_str = ''
all_json = []


class article(xml.sax.ContentHandler):
    def __init__(self):
        self.CurrentData = ""
        self.author = ""
        self.title = ""
        self.pages = ""
        self.journal = ""
        self.authorIndex = 0
        self.authors = []
        self.authorNum = 0

    # 元素开始事件处理
    def startElement(self, tag, attributes):
        self.CurrentData = tag
        target = ['incollection',"article",'mastersthesis']
        if tag in target:
            if len(self.authors) is not 0: