【作者】 赵海雁；

【导师】 陈立潮；

【作者基本信息】 华北工学院 ， 计算机应用技术， 2004， 硕士

【摘要】 随着网络技术的不断发展，特别是信息技术的发展，使社会进入了一个信息爆炸的时代，世界也由于Internet而变小，人们通过各种技术，如电子邮件、网络视频会议系统、网络传呼机等工具，使得交流突破了空间的限制。在网络技术的发展与普及的同时网络安全问题引起人们的广泛关注，研究并开发出一种能有效地实时拦截和监视可疑的网络信息系统，具有重要意义，基于这一情况，本文针对网络数据的捕获与信息过滤等问题进行了深入的探讨和研究。本文将信息过滤中的相关技术引入到网络安全领域，设计了网络数据捕获和信息过滤系统的总体方案与实现的方法。通过利用WinPcap开发包，实现了网络数据包的捕获，利用TCP/IP协议的封装理论，通过自下而上的分析，实现了地址过滤、类型过滤、协议过滤与端口过滤，通过对SMTP协议、POP3协议通信分析，提取了邮件，并通过过滤系统对邮件进行过滤，通过对HTTP协议分析，实现基于目标地址的访问控制。本文设计了中文分词处理算法，此算法通过对停用词库、单字词库与前置词库进行匹配，并通过对分词进行切分，利用改造型马尔可夫N元语言模型处理分词中出现的歧义词方法，建立了中文分词词库，然后对分词进行权重评价从而提取特征向量，并作为规则归入规则库里，作为文本过滤的标准，最后提出防止敏感信息通过的一些过滤策略。更多还原

【Abstract】 With the gradual development of the network technology. especially the development of the information technology, an information-absence era is becoming an information-explosion era. The world is becoming smaller and smaller just because of the internet. People can break the limitation of the space just through every kind of technology, such as e-mail, network video meeting system, network BP calls and so on and so forth. At the same time of the development and popularity of the network technology, the network security is becoming one of the people’s great concerns. Therefore, it is of great meaning to study and develop a new kind of network information system which can effectively and timely monitor and intercept the suspicions information. Based on the above situation , the article chiefly studies the date catching and the information filtering of the network.By analysing the general process of the information filtering, somerelative technologies in the information filtering are introduced into the field of network security, and the general scheme and methods for the realization of the network data filtering and the information filtering system are designed and achieved. The network data packets can be captured by means of WinPcap. According to the basic principle of TCP/IP and the analysis results, the SMTP protocol filtering,the port filtering,the address filtering and the kinds filtering can be achieved. E-mails can beattained by the analysis of the SMTP protocol and POP3 protocol. The control on the visiting target address can be implemented by using the filtering system on the e-mails and the analysis on the HTTP protocols.And also in this article, the Chinese word automatic segmentation is designed which uses the melio-rative Markoff statistics process. The Chinese word bank is built up by matching the unimportant-word base, the single-word base and the first-word base. The feature vector is drawn out through assessing the weight of every word, enters into the regular band as regulations and becomes a criterion for text filtering. <WP=6>And finally the article puts forward some preventive strategies to the filtered unhealthy information in order to make the local network secure.更多还原