Research And Design Of Personalized Agriculture Search Platform

In recent years, the development of vertical search engines has overcome some weaknesses of traditional search engine. But now the vertical search engines are still based on keyword match, unable to solve the problems of the keyword match method; In the agricultural field,the vertical search engines face the same problems, in the support of key projects in the national science &technology pillar program,Agricultural vertical search engine based on ontology (2006BAD10A050 2),this paper designs the personalized platform based on the agriculture search engine.At one hand,this paper studies the basis of personalized platform that associating the agriculture information with the geographic information.In this paper we design the agricultural market name geo-positioning system.This paper designs three methods to enhance the ability of recognizing the market name,and there are the geographic market positioning based on geographic information database, the geographic market positioning based on the search results of search engine, the geographic market positioning based on the number of search results of search engine.At the other hand this paper studies the method of creating personalized agriculture search platform.The information which the users are interested in are mined from the server logs,then the system expands the information semantically based on ontology to construct user interest model. When users retrieve relevant information, the system can expand the keywords based on user model,then calculate the similarity with the documents,and output the sorted results.At the beginning,the paper introduces the basic knowledges of the search engine and ontology.then the paper describes the characteristics of the names of the agricultural market which crawled from the Internet.then we build the multi-level geographic information database which can recogonize the alias of the cities in accordance with the characteristics of the market name. we analyse the possibility of the market name and the city which the market locates being at the same page on the Internet,then we give the method of the geographic market positioning based on the search results of search engine.The system can evaluate the realationship of the market name and the city where the market locates based on Pearson Chi-Square.then the paper studies the method of building and updating the user interest library.Firstly,the system builds the initial user interst library according to registration information of the user,then mines the server log and calculates the response coefficient to modify and update the library.And then the paper studies the method of caculating the user interest rate of the keyword according to the behavior on browsing the site of the user.Then the system expands the vocabularies in the user interest library semantically based on ontology ,and constructs the user interest model,then updates the model through caculating the Interest Rates of the vocabularies. When users search the information about the keywords, the system can expand the keywords based on the created user interest model,and provide the personalized services based on user interest model. Finally, a summary of this paper is given, and some problems needing further study are discussed.