Design And Implementation Of User-customized Desktop Search Engine

With the explosive growth of Internet information, the personalized needs of usersis increasing. In practical scenario, user often wants to do customized search in specificfields. Focused crawler is the base of customized search. The traditional focused crawlstrategy needs the support of the corresponding topic of train set. They can not workwith train set. Therefore, traditional vertical search engines only provide search servicefor one or some topics. They can not fully meet the personalized needs of users.The research of this paper is work for solving these problems. The maincontribution of this paper are described as follows：Proposed a web page extract algorithm for adaptive focused crawling. Adaptivefocused crawling need to extract the topic text of web page accurately. The existingmethod can work well with topic-type web pages, but they did not consider how toextract topic text from hub-type web pages. The algorithm proposed by this paper is ageneric extract algorithm(EATI), it work well in focused crawling.Proposed the adaptive focused crawl strategy. This strategy does not depend on thetopic train set. It work with a set of initial urls with same topic. With the same initialurls, adaptive focused crawl strategy can achieve the harvest rate level of best fisrtstratey which use a topic train set.Design and implement the user-customized desktop search system. This systemprovides a crawler to user which can be customized flexibly. User can do customizedcrawling by configuring the crawler simply to meet their needs. The system index theweb pages automatically in background and provide search service for users.