基于改进的模糊C-均值聚类的信任文摘  

Trust Summarization Based on Improved Fuzzy C-Means Clustering

作　　者：张泉[1,2] 曾国荪[1,2] 王伟[1,2] 孙明军[1,2] 谷华楠[1,2] 

机构地区：[1]同济大学计算机科学与工程系,上海201804 [2]同济大学嵌入式系统与服务计算教育部重点实验室,上海201804

出　　处：《计算机研究与发展》2008年第z1期268-273,共6页Journal of Computer Research and Development

基　　金：国家“八六三”高技术研究发展计划基金项目(2007AA01Z425);国家“九七三”重点基础研究发展规划基金项目(2007CB316502);国家自然科学基金项目(60673157)

摘　　要：信息时代的到来和互联网的发展,使信息文本呈爆炸趋势生成和传播,虚假信息的大量存在,给人们高效地获取可信的、安全的信息带来了相当的困难.如何对互联网上的信息文本进行信任评估,是内容信任和网络安全急待解决的问题.借鉴传统的自动摘要技术,首先提出了信任文摘的概念,在文本的词、句子、篇章等各个层面上发掘信任信息,改进自动分词方法,选取信任中心句并运用改进的模糊C均值聚类算法对其聚类,然后为信任中心句选择信任支撑句,最后生成了信任文摘,为基于内容的信任评估提供了一个较好的手段.The coming of information age and the development of Internet make information texts generate and broadcast explosively. A lot of mendacious existing information texts bring many difficulties for people to acquire trusted and security information. How to make a trust evaluation for information texts from Internet is a pressing problem of trust content and network security. In order to solve the problem, we refer the conditional automatic summarization technology, first we introduce the meaning of trust summarization. We explore trust information on every layer of text such as words, sentences and chapters, and improve word segmentation method. We choose the initial set of trust-centered sentences and cluster them based on improved C-Means clustering, and then we select trust-supporting sentences for every trust-centered sentence. Finally we generate trust summarization which is a good means of trust evaluation based on content.

关 键 词：信任文摘 内容信任 模糊C-均值聚类 自动文摘 

分 类 号：TP309[自动化与计算机技术—计算机系统结构] TP393.08[自动化与计算机技术—计算机科学与技术]