基于KPS的Web信息抽取MAS模型的研究  

Research on Web Information Extraction MAS Model Based on KPS

作　　者：钱君[1] 段隆振[1] 熊必成[2] 张和江[3] 

机构地区：[1]南昌大学信息工程学院计算机系,江西南昌330031 [2]江西旅游商贸职业学院,江西南昌330100 [3]苏州大学体育学院,江苏苏州215021

出　　处：《计算机与现代化》2007年第9期79-82,85,共5页Computer and Modernization

基　　金：江西省重点攻关项目(20061B01002)

摘　　要：在现有的相关研究的基础上,提出一个基于KPS的Web信息抽取的多Agent系统(MAS)模型。基于KPS的Web信息抽取方法集合了分别基于关键字、模式和样本的信息抽取技术的优点,有利于满足抽取高精度、高复杂度和高性能信息的要求。一个多Agent系统(MAS)可以把问题由大化小,由复杂变简单,每个Agent完成一个小目标,Agent间通过协调和合作,共同完成复杂的Web信息处理。而将基于KPS的Web信息抽取方法与多Agent系统进行有机的结合,可以大大降低Web信息抽取的难度,提高信息抽取的精度,从而更大地满足用户的需求。Based on the existing related research,this article proposes a Multi-Agent System model for Web irformation extraction based on KPS. 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 demands for extracting the information 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 com- plex Web information. Combining Web information extraction with Multi-Agent System can reduce the difficulty of Web informa- tion extraction,enhance the precision of information extraction, and then meet users＇ demands to more extent.

关 键 词：WEB信息抽取 KPS MAS AGENT 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]