Towards the Semantic Web: Collaborative Tag Suggestion

Content organization over the Internet went through several interesting phases of evolution: from structured di rectories to unstructured Web search engines and more recently, to tagging as a way for aggregating information, a step toward s the semantic web vision. Tagging allows ranking and dat a organization to directly utilize inputs from end us ers, enabling machine processing of Web content. Since tags are c reated by individual users in a free form, one important prob lem facing tagging is to identify most appropriate tags, while eliminating noise and spam. For this purpose, we define a set o f general criteria for a good tagging system. These criteria include high coverage of multiple facets to ensure good recall, least effort to reduce the cost involved in browsing, and high popu larity to ensure tag quality. We propose a collaborative tag suggestion algorithm using these criteria to spot high-quality tags. The proposed algorithm employs a goodness measure for t ags derived from collective user authorities to combat spam. Th e goodness measure is iteratively adjusted by a reward-penalty algorithm, which also incorporates other sources of tags, e.g. , content-based auto-generated tags. Our experiments based on My We b 2.0 show that the algorithm is effective.