Research On Entity Search Technology Based On Domain

As an important method for users to get information from Web,the search engine has already developed maturely,which can meet people's growing demand for searching.But the large search engines process the query with the method of keyword matching,this way leads to the loss of semantic information of the original search statement,which makes it unable to accurately understand the user's search requirements;usually,the form of results returned to the user is the list of web pages' title and a brief summary,so users need to look for the desired result information in each page,this whole process would reduce the users' search efficiency.In view of the above problems,this paper studies the related technology of entity search,and the main contributions are as follows:(1)Proposes an entities search method based on encyclopedia knowledge base.Based on the analysis of the user's search intention,we analyze the search statement,and define the semantic components as kernel concept and modifiers,and give a detailed analysis strategy.The data characteristics of encyclopedia knowledge base are analyzed,and the search statement are reorganized according to the semantic blocks parsed from the search statement.It needs to select the most effective sub-search statement which has the corresponding encyclopedia page,then select candidate entities in encyclopedia pages,and compare the original search statement's modifiers with candidate entities' attributes,according to the result of judgment,the correct entity results can be selected and return to the user.(2)As the fact that there are no corresponding encyclopedic pages for some entities,we propose the entity search method based on entity characteristic.This method sets the research scope on the Web page,and integrates domain oriented web data to local database.On this basis,a conceptual model of entity characteristic is defined,and the local characteristic library is built.We study and propose the method of how to make sure the related sets of web pages according to the analysis of search statement,and analyze strategy to achieve result entities by screening web content according to the entity characteristics,and the feasibility of the method is verified by experiment.(3)The prototype system is implemented according to the entity search method based on entity characteristic proposed in this paper.The system is based on the college entrance area,and web crawler has been achieved according to the data characteristics of the entrance area which aims to make information integration of web pages of colleges to the local database,and build an entity characteristic library in the field of the college entrance examination.Based on the above data set,the entity search function of the prototype system is realized,which verifies the effectiveness of the proposed method in this paper.