Automatic Classification of Social Tags with Thesaurus and Syntax Rules
Structuralized social tags can help improve the information organization and retrieval performances in social tagging systems.However, prior studies either applied manual approaches for tag structure analysis or employed automated tag classification methods based on thesaurus or syntax rules.To build a scalable yet reliable tag analysis framework, we combine thesaurus with syntax rule for automatic classification of tag types.Specifically, built on top of the prior studies, tags are firstly divided into three basic types factual, subjective and personal.Then, we construct a thesaurus based on the metadata obtained from an online wiki and a social tagging system.At the same time, we also design a list of rules based on the syntactic structure of social tags.Finally, the rules and thesaurus are combined in one unified framework.The effectiveness of the proposed framework is evaluated based on a large-scale dataset with 675 351 Douban movies and their related social tags.The results suggest that the proposed method can be effectively used for automatic tag classification.
tag classification self-built thesaurus syntactic rules
ZENG Ziming ZHOU Zhi QIN Siqi
School of Information Management, Wuhan University, Wuhan 430072, China
国际会议
武汉
英文
40-44
2016-10-21（万方平台首次上网日期，不代表论文的发表时间）