企业搜索引擎个性化排序方法

摘要

Traditional search engine is retrieval-oriented rather than user-oriented,in view of this weakness,the paper introduces the idea of personalized service to the enterprise search engine ranking.The user interest modeling technology which is the core of personalized service was researched in the paper.We apply the user interest model to query expansion and ranking and designed the personalized ranking method for the enterprise search engine.So the engine can display different result of the same retrieval for different users.We apply the personalized ranking to enterprise search engine in oilfield enterprise search and the tests show that,the research not only can effectively enhances the retrieval accuracy of the enterprise search engine and satisfies the user's personalized retrieval demand,but also has good adaptability.%针对传统搜索引擎“面向检索”而非“面向用户”的缺点,将个性化服务思想引入到企业搜索引擎排序中,对其关键技术即用户兴趣建模进行了研究,将模型用于查询扩展及排序中,并为企业搜索引擎设计基于用户兴趣的个性化排序方法,能为不同用户的同一检索请求提供不同的检索结果列表.通过将研究用于油田企业搜索引擎的实验证明,本研究能有效地提高企业搜索引擎检索精确度及满足用户的个性化检索需求,并具有较好的自适应能力.