简介：本文通过将模式识别中的K近邻算法和K-均值算法融合在一起,将其运用到入侵检测领域中,使它能够适应入侵检测的需要,通过实验分析表明,在运用结合之后的算法后,系统不仅能够保证实时性,并且具有了一定的未知入侵检测能力。

简介：ThekeyforriceplantsurvivalunderNaClsaltstressismaintainingahighK+/Na+ratioinitscells.Selectionforsalttolerancericegenotypesbasedonphenotypicperformancealonewilldelayinprogressinbreeding.Useofmolecularmarkersintandemwithphysiologicalstudieswillhelpinbetteridentificationofsalttolerantriceaccessions.EightriceaccessionsalongwiththecheckDongjinwerescreenedusing1/2Yoshidasolutionwith50mmol/LNaClattheseedlingstage.TheaccessionsIT001158,IT246674,IT260533andIT291341wereclassifiedassalttolerantbasedontheirK+/Na+ratios.SeventeenSSRmarkersreportedtobeassociatedwithK+/Na+ratiowereusedtoscreentheaccessions.FiveSSRmarkers(RM8053,RM345,RM318,RM253andRM7075)coulddifferentiateaccessionsclassifiedbasedontheirK+/Na+ratios.BandingpatternoftheaccessionswasscoredcomparedtothebandingpatternofDongjin.ThestudydifferentiatedaccessionsbasedontheirassociationofK+/Na+ratiowithmolecularmarkerswhichareveryreliable.Thesemarkerscanplayasignificantroleinscreeninglargesetofricegermplasmsforsalttoleranceandalsohelpinidentificationofhigh-yieldingvarietieswithbettersalttolerance.Thesalttolerantaccessionscanbetakenforwardintodevelopingbettervarietiesbyconventionalbreedingandexploringgenesforsalttolerance.