基于R-tree和倒排文件的混合索引的设计与实现

Mobile user always expect the search engine can return web pages whose content are not only matched with the query keywords, but also closed to user's location. However, the existing search engines only take the text relevance into account when give out the result web pages, and ignore the geographical proximity between the user's location and the location focus of the web pages. This paper is just proposed to resolve this problem. The paper presents a naive hybrid index structure based on R-tree and inverted file. Through the design and implementation of the hybrid index structure, the end user can conduct a real mobile search, and get web pages both have text relevance with the query keywors and close to the user's location. In addition, the paper analyses the result pages get from the hybrid index and traditional search engines, and verifies the high accuracy and efficiency of the hybrid index.