The Integration Of Folksonomy And Thesaurus

Folksonomy system is a civilian classification, whose classification tags have the advantage of individual spontaneous definition and classification openly-sharing, but have the disadvantage of semantic ambiguity, inaccuracy, indiscretion, which leads to the low efficiency of user information discovery and share. Additionally, the individual tag spells errors, folksonomy in Chinese segmentation structure ambiguity and the obvious difference of semantic cognition in different language context need to be resolved. Facing the development bottleneck of folksonomy system, we resort to the traditional classification. The traditional taxonomy centered on the subject, after years of development and practical application, its architecture has been very mature and perfect. Using the traditional taxonomy, we optimize and improve folksonomy, so as to improve the quality and efficiency of the organization and classification of network information, that is, to be able to apply the semantic relations of Chinese Thesaurus to extend user’s tags, to play a role of recommended and retrieval tags, to analyse user tag data in real time, so as a data source of thesaurus vocabulary update.Firstly, taking high-frequency subject words "Classified Chinese Thesaurus"(education), Del.icio.us data(users, tags, resources) as the data source, this paper are analyses the characteristics of filtered Chinese tags and subject words, explores the possibility of integration of Folksonomy and controlled Thesaurus. Secondly, the paper has built tag vector by the resources which were tagging by tag, has built co-occurrence matrix of tags and has built similarity matrix of tags, has clustered tags by SPSS software. The paper has built a small "tags tree"(the tags hierarchy) combining the tag of the similarity coefficient with tags clustering structure. At the same time, with the help of online Thesaurus "Classified Chinese Thesaurus" and ERIC thesaurus, we built light tag ontology. To use light tag ontology we extended the "Classified Chinese Thesaurus". Lastly, we have obtained25high-frequency tags as "extended subject words" of "Classified Chinese Thesaurus" through experiments, which has verified the effectiveness of the algorithm.The innovation of this paper lies in:on the one hand, we has designed a set of process of building light tag ontology based on online thesaurus, and the process was verified by the education class tags; on the other hand, we has proposed a algorithm to expand controlled vocabulary based on the built light tag ontology, and has verified the validity of the algorithm.