Internet Traffic Time Series

《Internet Traffic Time Series》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、ANFISMethodforForecastingInternetTrafficTimeSeriesSamiraChabaaAbdelouhabZeroual,JilaliAntariDepartmentofPhysicsDepartmentofPhysicsCadiAyyadUniversity,FacultyofSciencesSemlaliaCadiAyyadUniversity,FacultyofSciencesSemlalia,BP2390,Marrakesh40001,MoroccoBP2390,Marrakesh40001,MoroccoEmail:s.c

2、habaa@ucam.ac.maAbstract-InThispaperwehaveappliedtheadaptiveneuro•e.g.short-andlong-rangedependence,self-similarityinlarge•fuzzyinferencesystem(ANFIS)whichisrealizedbyantimescaleandmultifractalinsmall-timescale[9].appropriatecombinationoffuzzysystemsandneuralnetworksforforecastingasetofi

3、nputandoutputdataofinternettrafficInthispaper,weappliedtheadaptiveneuro-fuzzyinferencetimeseries.Severalstatisticalcriteriaareappliedtoprovidethesystem(ANFIS)methodformodelinginternettraffictimeeffectivenessofthismodel.Theobtainedresultsdemonstratethatseries.theANFISmodelpresentagoodprec

4、isioninthepredictionprocessofinternettrafficintermsofstatisticalindicators.ThisII.ADAPTIVENEURO-FuzzyINFERENCESYSTEMSmodelfitswellrealdataandprovidesaneffectivedescriptionof(ANFIS)networkconditionatdifferenttimes.Systemmodelingbasedonconventionaltools(equationKeywords-ANFIS,timeseries,in

5、ternettraffic,modelingdifferential)isnotwellsuitedfordealingwithilldefinedanduncertainsystems.Bycontrast,afuzzyinferencesystememployingfuzzyif-thenrulescanmodelthequalitativeaspectsI.INTRODUCTIONofhumanknowledgeandreasoningprocesseswithoutNetworksanddistributedprocessingsystemshavebecome

6、employingprecisequantitativeanalyses[10].Theadaptiveanimportantsubstrateofmodeminformationtechnology[1].neuro-fuzzyinferencesystem(ANFIS)isrealizedbyanThesesystemsknowarapidgrowingthroughouttheworkappropriatecombinationoffuzzysystemsandneuralnetworks.placewhichhasgivenrisetoadiscontinuit

7、yinexpertiseofThishybridcombinationenablestoutilizeboththeverbalandhumanoperatorstomanagethem.numericpowerofintelligentsystems.Commonly,fromtheSeveralresearchhavedemonstratethattheactualnetworktheoryoffuzzysystems,differentfuzzificationandtrafficpresentsdynamiccharacteris