面向信息内容安全的文本过滤和分类系统研究与实现

Research and Implementation of Text Filter and Categorization System for Content Security

本文设计并实现了一个面向信息内容安全应用的文本过滤与自动分类系统。系统采取探测器和分类器两个步骤实现高速网络环境下数据截取、还原、分类的功能。探测器采取简单规则匹配和高速字符串匹配算法来提高数据截取和自身的过滤性能,分类器采取基于简单向量空间模型设计,采取自动学习和人工干预相结合的方法来提高系统的查准率和查全率。本文给出了探测器和分类器在实际应用时的技术性能。

A text filter and categorization system is designed and implemented for content security application. We use Detector and Categorizer to capture, reassemble and filter packet stream in high traffic network. The Detector only use simple rules to filter packets. Combining automatic learning and manual processing, The Categorizer use the simple SVM to categorize. The evaluations and results is given.