基于活跃度指数的标签相关性判断研究

Study on Social Tags Relevance Judgment Based on Activity Index

[目的/意义]借鉴活跃度指数的设计思想,提出一种新的标签相关性判断策略,以改善标签相关性判断的效果和策略的通用性。[方法/过程]结合标签相关性判断的特点对活跃度指数的计算方法进行改造,进而提出一种基于多次活跃度指数迭代的标签相关性判断策略,并以社会化标注社区"豆瓣电影"的67 5351位用户的标签数据为例进行实验,以验证策略的效果。[结果/结论]实验结果显示,该策略的召回率为79.6%,准确率为93.3%,均较为理想,明显优于常用的Top-N策略。同时,该策略的通用性较好,适用于视频、音频、文本等各类型媒体。因此,该策略能够较好地解决标签的相关性判断问题。

[ Purpose/significance]The purpose of this paper is to improve both the effect and generality of social tags relevance judgment strategies. [ Method/process] On the basis of adjusting the calculation method of activity index by considering the characters of social tag relevance judgment, this paper proposes a tag relevance judgment algorithm based on multiple iteration of activity index, and verifies it by experiment based on 675351 Douban Movie users＇ social tagging data. [ Result/conclusion] The results show that this tactic performs well whose recall is 79.6% and precision is 93.3 % , and it is obviously better than Top-N social tags relevance judgment strategy. Moreover, it has good universality, and is suitable for various media such as video, audio, txt etc. Thus, it can be concluded that this tactic can solve the problem of tag relevance judgment well.