對社交標籤系統提供瀏覽式查詢之技術研究

The recent studies have showed that it can improve researching effectiveness by applying tags of objects in a social tagging system. By using the keyword based search interfaces, most users do not use advanced search functionalities and the average query length is very short. Therefore, it is not easy to completely describe the information need and search context of users according to the inputted queries and the amount of the returned results is usually huge. Although current web search engines have provided strong functions for ranking the related results, it is not necessarily that the ranking results satisfy the information needs of users. The users have to spend much time to browse the results one by one for getting the required data. Besides, a flat ranked list of the returned results makes it very difficult for the user to explore and find important items relevant to their search needs. For this reason, the goal of this research project is to study how to combine the existing information retrieval and data mining strategies to automatically discover representative tags and generate hierarchical faceted categories of tags from the search results. The discovered information can help users explore the returned search results and further describe their information needs more clearly for getting the required data items effectively and efficiently. The research issues involved in this project consist of the following topics: 1) The method for finding representative tags of query results, 2) The method for discovering representative semantic tag sets of query results, 3) The indexing and query processing for semantic tag sets, and 4) Facet hierarchy construction for tags of query results. The execution period of this project is two years. Not only theoretical methods for each research topic are studied, but also the corresponding performance evaluations are implemented to validate the effectiveness of the proposed strategies.