基于Agent的智能元搜索引擎技术研究

Research on Intelligent Metasearch Engine Based on Agent

针对现有搜索引擎存在的问题,提出基于Multi-agent的分布式搜索引擎系统。系统采用元搜索引擎结构,利用Agent技术和基于个性化模式的信息过滤技术,使系统具有一定的智能性。通过个性化检索和分类浏览相结合的检索方式可提高搜索结果的可浏览性。结合数据库的分类和虚拟语言模型方法实现了资源选择的优化。提出基于文本/位置分析和群决策的合并算法,对搜索结果的标题和文档片断信息进行相关度分析,将文本分析与规范化的搜索结果位置信息相结合,计算文档的相关分值,最后采用基于群决策的合成方法对搜索结果进行一致性排序。试验结果表明,提出的元搜索系统具有较好的搜索效果。

A intelligent metasearch engine system based on multi-agent is proposed. The agent technique and information filtering technique based on personalized models are utilized, which makes the system more inteligent. The retrieval methods combining customized search with classified browse help users find relevant results more quickly. The scheduling of search sources is optimized by integrating the database categorization with virtual language model approach. A result merging method based on text / rank analysis and group decision making activity is presented. By utilizing textbased information such as title and snippets obtained from search results, the method to analyze the relevancy of title and snippets is described. Then, the relevant scores of the relevant documents are normalized by incorporating text anal- ysis together with rank. Finally, a merging method based on group decision making activity is adopted to sort the search results. The experimental results show that this system has a better performance.