Research On Tag Generation And Ranking Algorithms

In recent years, the Internet is flooded with massive text information. In order to strengthen the management and use of information, the industry introduced the tool of Tag. The tag is a new kind of information description form, it can be roughly interpreted as keywords or subjects, and it is a brief summary of the text. Use tag can help us easily discovery, management and disseminate information.The first step to use tag is obtain tags, the industry mainly uses two kinds of way to get tags:use data mining and machine learning algorithms to automatically generate tags from text and use crowd-sourcing system to invite users donate tags. This paper summarizes the popular tag generation methods in the industry.Furthermore, in order to improve the quality of the tag and make deeper use of the tag, we also studied the existing tag sorting algorithms, and then proposed a new tag ranking algorithm called CBTR. The algorithm is based on user’s tag-click behaviors and probabilistic model.Finally, we proposed a new tag-based book search system and apply it in CADAL digital library. Tag-based book search system can effectively identify the subject of the book. And it can better solve the problems in book search system:lack of text information, multi-dimensional search and too much search restrictions. Tag-based book search system greatly improved the capabilities of search and management digital resources.