Intheearlydays,outlierswereconsideredanomalous,possiblyerroneousobservations,whichshouldbeidentifiedandremovedfromtheanalysis,whenbuildingthemodel.However,analyzingsuchoutliersmightbeusefulinsomecase.Inthispaper,wedeveloparecommendersystemtoidentifyandexplainab-normalaccessestooursystem,theOnlineLearningSystembasedonMoodle(OLS)andtheonlineEducationalManagementSystem(EMS).Ourapproachisproposedonthebasicofone-classclassificationandSupportVectorMachinetoprocessandclassifyrawtextdata.Ourexperimentsdemonstratetheutilityandaccuracyofthesystemindatasetfromtheusageofoursystem.