摘要
随着移动互联网的飞速发展以及社会化网络的出现,微信等移动社交类应用在生活的各个层面为人们带来便利,但同时也带来了谣言、泄密等网络安全问题。传统的基于字符串匹配的关键词过滤审计引擎已经不能满足人们的需求。针对传统特征项权重算法在处理变异短文本方面的不足,提出了改进的TF-IDF算法,同时将社会网络计算的相关技术应用到审计引擎中,实现了基于内容的事件抽取和表达。实验结果表明,改进后的算法具有较高的准确率和召回率。
With the rapid development of Mobile Internet and Social Networking,convenience is brought by mobile social networking applications,such as WeChat,in all aspects.Thus rumors,leaks and other network security issues are brought as well as convenient.However,the traditional audit engine based on string matching and keyword filtering didn＇t meet people＇s needs any more.The improved TF-IDF algorithm is proposed based on its inadequate in dealing with abnormal short texts.The relevant social network computing technologies are applied in audit engine to realize the analysis of event situation based on content.Result shows that the improved algorithm has higher recall rate and accurate rate,compared with the traditional TF-IDF algorithm.
作者
饶洁
RAO Jie （School of Electronics and Information Engineering, Tongji University, Shanghai 201804, China）
出处
《电脑知识与技术》
2014年第4期2423-2427,共5页
Computer Knowledge and Technology
