Research On Framework And Model Of Web Search And Expert Search Engine

Along with the development of the network, the development of search engine has been in the advance. And web search engine increasingly has become the most important information retrieval tool and also a critical network application in the network. The framework and the information retrieval model of the search engine are core technology.This thesis aims to do some research and modification on the framework and retrieval model of search engine, and then apply the fruit to the expert retrieval system. Ultimately, such work makes sense to the beginning research of the object-based search engine.In this paper, the major work is in such aspects below: First, introduce the framework and retrieval model of the full-text search engine which is the most widely used. And then do evaluation experiment on the the most classic and most commonly used retrieval model. Based on the result of the experiment, the writer studied the performance of the different statistic language models which is based on different soomth way using differet values of the parameter. Second, based on requirements and characteristics of the expert search task, writer used the two-stage retrieval model and expert window model as solutions to complete the expert search task. Two-stage retrieval model is a new retrieval method which is used to deal with the new search task. It is based on the using of general retrieval models and is an attempt to the modification of the retrieval process .Of course, such an attempt inherits the general retrieval model, because writer organized the the general retrieval models into two-stage expert retrieval model in the optimized order based on specific circumstance.Then we get good performance in the experiment using the two-stage model.Third, based on expert window model which is the second method for the expert retrieval system, writer also did a amount of experiments .Then compared with the results of the experiments done by the two-stage model, writer select the best one as model of the Campus Object Search Engine.Fourth, writer used retrieval model of expert search engine which gets the best performance as the core model of Campus Object Search Engine. The Campus Object Search Engine can not only return the documents relevant with the queries as search results, but also return relevant objects which are extracted from the corpus. Namely, it is structural oinformationAt last, the writer summe the article and pointed out that the issue needs further in-depth research.