一种启发式的信息抽取算法  

A Heuristic Information Extraction Algorithm

作　　者：吴芬芬[1] 刘磊[1] 肖宪[1] 

机构地区：[1]吉林大学计算机科学与技术学院

出　　处：《吉林大学学报（理学版）》2007年第1期73-76,共4页Journal of Jilin University:Science Edition

基　　金：吉林省科技发展计划项目基金(批准号:20050527)

摘　　要：提出一种启发式的信息抽取算法,并利用该算法建立一个信息抽取系统.该系统基于文本分块,利用文本的语义特征和结构特征,抽取具有特征的状态,在此基础上,利用反向动态规划和正向A*算法,抽取剩余的无特征状态.通过对100篇论文头部进行测试的结果表明,精确度和召回率比基于单词和传统Viterbi算法的方法均有所提高,启发式算法的性能优于Viterbi算法.A heuristic information extraction algorithm is presented and an information extraction system is built with it. The system utilizes the semanteme characteristic and structure characteristic of the text to make the states with certain characteristics. On the basis of this result, we carried out extracting the remainder states having no characteristic with a algorithm incorporating backward-dynamic-programming with forward-A^＊ algorithm. We have tested 100 pieces of headers of computer science papers provided by the search-engine research group from CMU university of USA. The result shows the recall and the precision rate are all improved a lot compared with existing methods which are based on words and traditional Viterbi algorithm. In condusion, the heuristic algorithm is better on performance than Viterbi algorithm.

关 键 词：启发式算法 文本分块 A^＊算法 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]