基于标引信息的网络新概念发现算法  

The Algorithm of New Conception Discovery Based on Indexing Information

作　　者：夏霙[1] 刘功申[2] 李翔[2] 

机构地区：[1]上海交通大学安泰管理学院 [2]上海交通大学信息安全工程学院

出　　处：《微型电脑应用》2007年第1期8-10,66,共3页Microcomputer Applications

基　　金：国家自然科学基金项目资助:项目编号60502032

摘　　要：网络新概念的发现和识别是信息安全领域的一个基础技术,它可以为网络信息智能处理、网络有效信息识别提供底层支持。通过合理地利用标引信息,算法在速度上可以满足网络海量信息的要求;为了提高新概念发现的准确率,算法不但采用了传统的分词、串频统计技术,而且还采用了字偏旁部首自动组合、网站间串频同现等新技术。实验证明,算法识别的潜在新概念中可接受率完全可以满足当前的应用需求。The discovery and recognition of network new conception is a fundamental technology in the field of information processing ,which can help sensing publicinterests in Internet and recognizing Internet interesting information. Because of using indexing information rationally, it can process a mass of Information with the algorithm proposed in this paper. In order to improve the precision of new conceptiondiscovery, the algorithm adopts not only the technology of traditional word segmentatio-nand string frequency statistics, but also automatic combination of Chinese character components and the string co-occurrence between websites. It's proved by experiments that 75 percent of the potential new conception is acceptable. The new algorithm can thoroughly meet the requirementsof the application of information processing.

关 键 词：新概念 串频统计 信息处理 网络挖掘 

分 类 号：TP393[自动化与计算机技术—计算机应用技术]