《基于层次分类的网络内容监管系统中串匹配算法的设计与实现》由会员分享，可在线阅读，更多相关《基于层次分类的网络内容监管系统中串匹配算法的设计与实现（74页珍藏版）》请在金锄头文库上搜索。
1、硕士论文 基于层次分类的网络内容监管系统中串匹配算法的设计与实现 I摘摘 要要 互联网上的反动、暴力、淫秽等非法信息使得网络内容监管日益重要。本文在对网络内容监管方面的研究现状进行分析后，采用一种层次分类的方法，即在信息识别过程中，先将网络信息进行关键字过滤，然后再将含指定关键字的信息内容传送给语义分析模块进行处理，从而减轻语义分析的压力，提高系统效率，并且适合于非法信息多变的特点。 关键字过滤通过串匹配算法实现，并且已成为系统性能的瓶颈。为此，本文分析了现有串匹配算法（包括单模式串匹配算法和多模式串匹配算法）的特点，设计并实现了一种适合于中文大字符集和网络内容监管的串匹配算法：基于 QS算法
2、的单模式串匹配算法（IQS 算法）和基于 Wu-Manber 算法的多模式串匹配算法（IWM 算法） 。IQS 算法主要用于工作人员的日后查阅，而 IWM 算法主要用于网络信息过滤。并且在中、英文两种环境下以及模式串的不同长度和个数情况下对IQS 算法和 IWM 算法进行了实验比较分析，结果表明在运行时间与尝试次数方面整体优于其它算法，获得了较好的效果。 IQS 算法和 IWM 算法已集成实现在网络内容监管系统中，用真实数据对该系统进行了测试，实验结果表明该算法表现出了较快的运行速度，系统的各项性能指标均已达到预定目标。 关键字：关键字：网络内容监管，串匹配，单模式，多模式 硕士论文 基于层次
3、分类的网络内容监管系统中串匹配算法的设计与实现 IIAbstract Network Content Monitor has become more and more important just because there is illeagl information about subversion, violence, obscenity etc. found on the network. In this paper, we analyse the current development of Network Content Monitor and introduce a method
4、 of hiberarchy classification for it. This method first filters the network information by keywords filter, and then with the help of semantic analysis analyses those information, which contain the keyword(s) given by the user. It can ease the pressure of semantic analysis, improve the efficiency of
5、 the system and adapt well to the caprice of illeagl information. Keywords filter is realized by string matching algorithm and is the performance bottleneck of Network Content Monitor indeed. Therefore, this paper analyses the existing algorithms of string matching including single-pattern string ma
6、tching and multi-pattern string matching. Then we design and implement two algorithms suitable to Chinese character set and Network Content Monitor: one is IQS, a single-pattern string matching algorithm based on QS algorithm, and the other is IWM, a multi-pattern string matching algorithm based on
7、Wu-Manber algorithm. The aims of the two algorithms mentioned above are to search a keyword among files saved in local hard disks and to filter the network information, respectively. Meanwhile, we make experiments on IQS algorithm and IWM algorithm under the circumstance of Chinese and English, and
8、different lengths and numbers of patterns. The experiments results indicate that the two algorithms have a better performance on speed and attempt. Finally, we integrate IQS algorithm and IWM algorithm into the system of Network Content Monitor, and make experiments on this system. The results show
9、that the high speed of string matching algorithms is achieved, and the performance targets of this system are also obtained. Key Words: Network Content Monitor, string matching, single-pattern, multi-pattern 硕士论文 基于层次分类的网络内容监管系统中串匹配算法的设计与实现 III目目 录录 摘 要. I Abstract.II 目 录.III 1 绪论.1 1.1 研究背景及意义.1 1.2 国内外现状分析.2 1.2.1 网络内容监管.2 1.2.2 串匹配.3 1.3 研究内容及目标.6 1.4 论文的安排.6 2 串匹配算法.7 2.1 概述.7 2.2 单模式串匹配.7 2.2.1 BF算法.7 2.2.2 KMP算法.8 2.2.3 BM算法 .10 2.2.4 KR算法.13 2.2.5 QS算法.
