| Summary: | 碩士 === 逢甲大學 === 資訊工程所 === 93 === With the innovation of network and information technology, there are more and more information that we can obtain on the Internet. For all of us, there are lots of ways to obtain knowledge right now. We can get knowledge not just surveying books or learning on the class, but also from computers, multimedia, and the Internet. By using the search engine, the World Wide Web becomes a large database that contains rich resources. Because of the characteristics of the Internet, we can find desired materials for study any time any where. For this reason, the learning habits of people are gradually changing. Although numerous searching engines for Web searches, there still has a lot of space for improvement. In this paper, we propose a flexible searching method by using data mining techniques. We analyze the theses from the specialized domain for helping us to get research papers closer to our needs. Even though a search engine can help us find lots of information, it sometimes contains lots of redundant data. Our method can solve this problem.
Our method uses two kinds of data mining techniques called association rule mining and negative association rule mining. We use association rule mining to find out the relationships between words hidden in the articles for discovering the useful word sets in order to help the user in searching documents. We also use negative association rule mining to find out the exclusive relations between words for filtering some useless document to get accurate searching results. Users can be used to use the feedback mechanism to rate the searching results. This can improve our method and meet users’ needs. We have developed the core of proposed method and examined it on the specialized Web sites like IEEE, ACM, and Google Scholar. The result shows that our method really can help users find the useful documents as needed.
|
|---|
