欢迎来到天天文库
浏览记录
ID:36515215
大小:553.49 KB
页数:74页
时间:2019-05-11
《物理隔离网闸的过滤技术研究和实现》由会员上传分享,免费在线阅读,更多相关内容在学术论文-天天文库。
1、上海交通大学硕士学位论文物理隔离网闸的过滤技术研究和实现姓名:邬雄崎申请学位级别:硕士专业:电子与通信工程指导教师:李生红20070601上海交通大学工程硕士学位论文摘要网闸的智能文本过滤的方案,来改善现有隔离网闸中的内容过滤模块的性能,并对其中的关键技术混合特征选择和KNN过滤算法的改进作详细地介绍。结合常规特征选择和粗糙集理论的混合特征选择方法,先利用传统的特征选择方法作为初步选择特征,然许再利用粗糙集进行二次特征选择使得特征项数目大大减少,去掉很多冗余属性,从而大大降低文本的维数;而改进的KNN过滤算法是结合了文本聚类技术和传统的KNN算法
2、,不仅有助于提高过滤效果而且也缩短过滤的处理时间,相比与传统的KNN算法,改进的KNN算法更适合于隔离网闸的应用。关键词:信息安全,物理隔离网闸,有限状态自动机,特征选择,粗糙集,文本过滤,向量空间模型,K个最近邻法第II页上海交通大学工程硕士学位论文ABSTRACTResearchesandImplementsofTextFilteringforPhysicalGAPABSTRACTWiththeInternetandE-GovernmentNetworkbeingwidelyusedinworkingarea,lifeaswellasstud
3、yarea,networksecurityisbeingconcernedincreasingly,securityofnetworkinformationisfocusedwhiledevelopingthenetworkinwholesociety,inventionanddevelopmentofPhysicalGapbasedontheisolationtechnologyisinaccordancewiththerequirementofnetworksecurity.Textfilteringbeingappliedcurrently
4、forPhysicalGapmostlyutilizeskeywordsfilteringandURLfilteringwhichhaspoorperformanceonprecessionandefficiencyforfilteringandwhichalsoarenotabletomeetthetoday’srequirementfornetworksecurity.Withresearchanddevelopmentfortextfilteringtechnology,especiallytheareafortextrepresentat
5、ionwithVSMandfilteringalgorithmbasedonVSMisbecomingmoreandmoremature,itishelpfulforustomakeoutaadvancedtextfilteringtechnologytoreplacekeywordsfilteringandURLfilteringcurrentlybeingusedforPhysicalGapinordertoimprovethetextfilteringperformanceforPhysicalGapandalsomeetmorestric
6、terrequirementofnetworksecurity.ResearchandimplementforIntelligentText第III页上海交通大学工程硕士学位论文ABSTRACTFilteringsolutionforSGAPisamaintaskofthispaper.Authorproposesonesolutionofintelligenttextfilteringbyusingsomerelevanttechniquessuchasfilteringalgorithmmodel,informationprocessing,
7、roughsettheoryandphysicalgap,thesolutionisbasedonstatisticsandsuitableforphysicalgaptoimprovecurrentkeywordsfilteringinSGAP.AdetailedintroductionforcriticaltechnologyinnewsolutionsuchashybridfeatureselectionandmodifiedKNNalgorithmismade.Hybridfeatureselectionistocombinetradit
8、ionaloneoffeatureselectionmethodswithroughsetattributereductmethod,p
此文档下载收益归作者所有