Performance evaluation and analysis of tag ranking and image search system using tag similarity

Abstract

The original tag-based systems use one-tag matchingfor search, and consequently cause insufficient accuracy. In addition, when users put tagson content, they do not consider relationship andorder of priority between tags, and therefore this tagsprovide inefficient information about contents. Previous research proposed cluster-based tag ranking and image search system using tag similarity. This research uses image search results’ accuracy and recall ratio to evaluate theproposed system. © Research India Publications.