User preference retrieval using semantic categorization for web search

Search engines have been one of the most popular ways for people to find web pages of interest. Presently, when a user enters a keyword in a search engine, the search results are usually presented the same result to other users who search the same keyword, which might not be related to each user's field of interest. Therefore, the researcher of this study would like to propose a new searching technique to get each user's most relevant information by using a user preference. This research will categorize user preference to build the user profile and general profile base on user's search history and category hierarchy, respectively. The search engines then use those profiles to determine the interests of each user, execute the search query to obtain a set of relevant documents, and reranking the documents in a manner that best reflects their relevance to the user's profile. Many algorithms have been designed, analyzed, implemented and experimented to find the most appropriate and the most effective one to create the relationship between each keyword and each category to best meet each user's preference.