Research And Designment Of Web Information Extraction MAS Model Based On KPS

With the very development of computer communication technology and the popularization of personal computer,as a network for the seven seas,Internet comes into more and more domains such as people's life,job,study and business.It is built between the great contents of Web information and the users' demands that bad imbalance and conflict,so people find it difficult to get their really interested data from a mass of information at a short time. And then the needs for acquisition technique on highefficiency information becomes more and more exigent to people. Through Web information extraction system,people can acquire their really interested information with high efficiency and also can finish the further step after information extraction,such as data mining and message retrieval,which can supply the possibility for making use of a mass of Web information again. So it's said that Web information extraction system owns obvious superiority and vast application prospects,and is the hot spot in so many researching domains nowadays.At present, system based on Agent becomes a very important content in computer science field. Therefore, Agent system fits for handling the problems which are all complicated, unpredictable; variational.The characteristics of internet also supply sufficient application foundation for Agent. A Multi-Agent System(MAS) is a network for problem solving which is made up of many Agents.In a Multi-Agent system, every Agent has its own knowledge base, aim and abilities.Many Agents work together for complex assignments to achieve a whole aim.This article will propose a Multi-Agent System model based on KPS for Web information extraction after studing the existed related researches.The method of Web information extraction based on KPS integrates respectively the advantages from the one based on keywords,one based on pattern and one based on sample pages,which is good for meeting the demand for extracting the infromation with high precision,high complexity and high performance. A Multi-Agent System can make the big problem become small and make the complex one become simple.Every Agent can achieve a small aim,and Agents can work together for handling complex Web information.That combining the Web information extraction with Multi -Agent System can reduce the difficulty of Web information extraction, enhance the precision of information extraction,and then meet users' demands to more extent.