Topic Hierarchy Generation Based on Anchor Text and Term-correlation

• Main Authors: Li-Ling Chou, 周麗玲
• Other Authors: Hahn-Ming Lee
• Format: Others
• Language: zh-TW
• Published: 2005
• Online Access: http://ndltd.ncl.edu.tw/handle/65841329574921528897

碩士 === 國立臺灣科技大學 === 資訊工程系 === 93 === As Internet booms prosperously, there is various information available for user to obtain, such as new technique information and course contents for instance. It has become an important task to provide "Topic Directory Query" Service in order to help users understanding relevant subtopics of their interested techniques within a short period of time. In this thesis, we propose an approach that utilizes Anchor Text and Term-correlation technique to construct and generate topic hierarchy, in order to facilitate users search effectively and efficiently the scope of their interested topics that differs from manually constructed topic hierarchy, such as Open Directory Project or Yahoo Web Directory for instance. In our experiment analysis results, our proposed approach was proved to be effective in searching relevant hierarchical subtopics, especially those topics that cannot be found from manually constructed topic directory search engine mentioned previously but can be found in our system. Therefore, with regard to "Topic Directory Query" Service, there are still many issues need to be resolved, such as precision rate enhancement and new topic detection. However, we still hope that learning new techniques for everyone will never be a troublesome problem. Furthermore, by promoting the concept of topic hierarchy generation, we hope the issues mentioned previously can be researched continuously.