基于微博用户行为的时区预测
首发时间:2011-12-16
摘要:本论文基于微博用户发表微博的时间数据建立了两个用户行为模型,即用户活跃度-时间模型及用户首条微博-时间模型,根据用户行为模型对用户数据进行统计。研究发现,不同时区的微博用户发表微博的高峰期到来时间不同,呈现出明显的时区特性。此外,本论文采用决策树预测算法,根据本论文所建立的用户行为模型,对微博用户进行时区预测,并给出各时区微博用户的预测结果。其中,采用用户活跃度-时间模型对东九区和东六区的用户进行预测,其预测结果的准确率为0.805,采用用户首条微博-时间模型对东九区和东六区的用户进行预测,其预测结果的准确率为0.76.
For information in English, please click here
Time Zone Prediction Based on User Behavior of Microblogging
Abstract:In this paper, two user behavior models which are user activity-time model and first weibo-time model have been built based on the time data. Statistics and study found that the peak time when user update weibos is different in every time zone. There is obvious time zone feature among users of microblogging. In the research, the decision tree prediction model has been used to predict the time zone of users. And prediction results are given out. The precision of prediction result based on user activity-time model between GMT+9 and GMT+6 time zones is 0.805. And the precision of prediction result based on user first weibo-time model between GMT+9 and GMT+6 time zones is 0.76.
Keywords: Time Zone Prediction Microblogging User Behavior Decision Tree
基金:
论文图表:
引用
No.****
同行评议
共计0人参与
勘误表
基于微博用户行为的时区预测
评论
全部评论0/1000