Development of Topic Navigation System for Assisting Searches on the Web

When we use search engines to investigate an unknown word and topic, more unknown words often appear on the Web pages of the search result. Consequently, we have to repeat the searches and read a lot of Web pages until we understand the word. This paper proposes a topic navigation system that visualizes the related words of an input word as a tree structure. The system also attaches a Web page explaining each word in the tree structure. The generated topic navigation informs us of other words that we should search before we start studying the input word and limits the Web pages needed to understand each word in the tree structure. Experiments were conducted to compare the topic navigation generated by our system and that generated by the Open Directory Project (ODP), which is a Web site with categories of topics found on the Web. The results of a questionnaire show that our system is more useful than ODP for understanding an input word.