基于Agent的智能化元搜索引擎个性化机制

Abstract

Information overload is a common issue in “big data” environments, because of which such environments require precise and intelligent retrieval tools. In this paper, we propose the design and implementation of a personalization function for an intelligent, agent-based meta-search engine that improves information retrieval efficiency in “big data” environments. We focus on individualized methods, along with the relevant theoretical and technological requirements for the implementation of a personalization function in an intelligent, agent-based meta-search engine, provide query analyses, and describe the interest mining and scheduling processes of the search engine. We then detail the recognition mechanism of complex inquires based on a dynamic learning process, the interest mining mechanism based on the general view model of user interest, which involves dynamic updates, and the strategy for search engine scheduling based on a concept lattice and user logs. Experiments showed that the personalization function for intelligent, agent-based meta search engines proposed here accurately identified complex queries, effectively learned users search interests, improved the relevance of the search results, and intelligently scheduled member search engine. It also helped improve the efficiency of information retrieval by users, thus enhancing their searching experiences.