Research And Implementation Of TheChinese Search Engine Based On Meta Search

Information retrieval is one of main purposes as we browse internet, there are some technical problem need to further improve in the two major Chinese search engines Google and Baidu, such as how to increase the precision for multi keyword enquiry, how to increase the recall by designing a Chinese Meta Search Engine (C.M.S.E.), how to shorten the search time, etc.In this paper, to increase the search precision mainly, we study and develop a C.M.S.E. with following approaches: (1) we propose a C.M.S.E. whole frame with three modules: enquiry agent, search agent and operation agent, their work flow are introduced too. (2) To choose the meta search engine's data sources, we compare the most popular ten Chinese search engines by recall and search time, and make choice of Google and Baidu as the basic data sources, theirs recall is up to 88.8%, this can well prepare for the design the high recall C.M.S.E. (3) In the single keyword information retrieval, we propose a new approach which the relevant arithmetic is based on abstract analyse, and present the percentage analytical method and ratio analytical method, the former calculate the precision with page title and its best result is 76.56%, the later do with page title and page abstract, and its precision is 72.74 %, comparing with Google and Baidu, the C.M.S.E. with ratio analytical methods can increase the precision by 3.16% and 6.21% respectively, it not only can perform the synchronous enquiry, but also save on a great deal of storage space and hardware requirements. (4) Considering the retrieval bug of the major search engines, Google and Baidu, we proposes a new "multi-level weight" method to improve the retrieval precision, the C.M.S.E. not only filters the dead-link and repeated-link, but also improves the precision by 12.37% and 18.05% respectively when comparing with Google and Baidu. (5)In C.M.S.E. system design and realization aspects, we introduce the database system, software system, and human-machine interface, it can complete the three functions, i.e. the multi keyword enquiry with high precision, single keyword timely enquiry, and the compare of precision and search time between the C.M.S.E. and Google & Baidu.