面向Folksonomy的用户兴趣相似性度量方法  

A users' interest similarity calculating method in Folksonomy

作　　者：张艳桃[1] 王国胤[1] 于洪[1] 

机构地区：[1]重庆邮电大学计算智能重庆市重点实验室,重庆400065

出　　处：《南京大学学报（自然科学版）》2013年第5期588-595,共8页Journal of Nanjing University（Natural Science）

基　　金：国家自然科学基金(61272060;61379114);重庆市自然科学基金重点项目(cstc2013jjB40003)

摘　　要：在社会化标签系统Folksonomy中,标签不仅能描述资源的内容,而且能体现用户的兴趣偏好.通过标签来表征用户的兴趣,定义了标注行为一致程度、用户资源共享程度和好友间兴趣相似度概念.使用用户对资源的标注行为一致程度来建立用户兴趣模型.通过统计实验发现:Folksonomy系统内好友间兴趣相似度高,但用户资源共享程度却较低;因此,将好友间兴趣相似度引入用户间兴趣相似度的计算公式中.将新的用户间兴趣相似度计算方法使用于SCAN社区发现算法中,社区发现结果验证了用户间兴趣相似性度量方法是有效的.Folksonomies are part of a new generation of tools for the retrieval,deployment, representation and production of information,commonly termed ＂Web 2.0＂. Tagging, which is one of the defining characteristics of Web 2.0 services, allows users to collectively classify and find information. A community of interest is a community of people who share a common interest or passion. It is important to discover the community of interest to help its members to take part in the community to exchange information, to obtain answers to personal questions or problems, to improve their understanding of a subject,and to share common passions or to play. The objective of this paper is to study a new measure method of similarity between users＇ in order to detect communities of interest. We introduce some concepts such as the degree of annotating consistency（DAC）, the degree of resource sharing between friends（DRSF）and the degree of interest similarity between friends（I）lSF）using tags to represent users＇ interest. We construct the user interest model through the degree of annotating consistency. According to results of the statistical experiments,we find that the DISF is higher but the DRSF is lower among friends in a Folksonomy system. Therefore, we improve the measure method of similarity between users＇ by introducing into the DISF. We revise the SCAN algorithm to detect community by using the new measure method,and the experimental results show that the proposed method is more effective.

关 键 词：FOLKSONOMY 兴趣相似性 社 区 好友关系 

分 类 号：TP393.4[自动化与计算机技术—计算机应用技术]