摘要
提出一种基于模块关系树的分析方法,考虑每个实体与用户之间的兴趣、住址和共同好友等相关因素,制定不同的关系树,然后根据路径长度计算各因素的相关度值,最后综合每个实体模块,从而筛选出关系最密切的实体。实验结果证明,该算法能过滤掉大量无关信息,有效找出最相关的实体,提高了搜索结果的准确率。
This paper proposed an analysis method based on the modular relational tree.It considered the correlative factor of interest,address and common factors between users and each entity,formulated different modular relational tree,then calculated the correlation of each factor according to the length of the path,and finally integrated module for each entity to filter out the most closely related entity.The experimental results show that the algorithm can filter out a lot of irrelevant information and identify the most relevant entity effectively,and the accuracy of search results is great improved.
出处
《计算机应用研究》
CSCD
北大核心
2012年第2期698-700,共3页
Application Research of Computers
关键词
兴趣关系树
地址关系树
共同好友
相关度
社交网络
信息过滤
interest relational tree
address relation tree
common friends
correlation
social network service
information filtering