英文PPT模板presentation

EpilepticSeizureRecognitionGENDAA01INTRODUCTIONTOEPILEPSY02DATAUNDERSTANDINGANDPREPARATION03MODELLINGPROCESS04STRENGTHS,LIMITATIONS&POTENTIALSGROUP6PRESENTATIONEpilepsy•65millionpeopleworldwidesufferfromepilepsy(theEpilepsyFoundation)•Approximatelyonein103peopleisaffected.Demography•Recurrentseizures•Lossofconsciousness•Damagetothebrain,orevendeath.(30min)Symptom(UniversityofOxfordandtheKarolinskaInstitute,Stockholm)PrematureDeath010203RateofPeoplediedprematurelyEpilesyPatients8.8%Others0.7%•Oneofthemostcommonneurologicaldiseases•Excessiveelectricalactivitywithinnetworksofneuronsinthebrain•Detection&Prediction•Allowforinterventionaltreatment•Improvecurrentepilepsydiagnosis•Medicalsignal-patternrecognitionframeworkEpilepsyEEGDataUnderstandingBrainActivityAnalogEEGSignalDigitalEEGSignalEpilepticSeizureRecognitionEEGSignalsDatasetStatistics•11,500records•5types•2,300recordseachtypeProperties•178Hzsamplingrate•1secondlongTypesofEEGWaveformsHealthy(TypeA)Healthy(TypeB)PotentialSeizure(TypeC)PotentialSeizure(TypeD)ActualSeizure(TypeE)DataPreparationDiscreteFourierTransformationTimeSeriesFrequencySeriesTimeDomainDatasetf(t)=𝐜𝐨𝐬(𝟓𝐭+𝛗)𝜑=π4𝜑=π𝜑=π2DataTransformationFrequencyDomainDatasetAdvantages•Removetheinitialtimeoffsets.•Extractmorerecognizablefeatures.•Delivermorerecognizabledataset.Limitation•Mightfilterusefulinformationoftheoriginaldataset.MODELLINGPROCESSSelectTop3AccurateModelsSelectbasedonmodelaccuracyinmodelrecalibrationModelSelectionTransformedDataReclassifyReclassifyDataTypessplitorcombineintocustomizedgroupsbasedonresearchpurposeBalanceSampleSizeofEachGroupEnsureevenpredictionabilityforeachgroupBalanceAutoClassifierComputer’sjob!!ModelsareconstructedandtestedforeachgroupModelConstructionPartitionDataRandomlyforDifferentUseinModellingAssignproportionofdatausedinmodelbuilding,recalibrateandtestforaccuratemodellingDataPartitionModelValidationValidateModelAnalyzetestingresultsofselectedmodelsRESULTANALYSIS&MODELVALIDATION95.29%95.39%93.45%94.60%80.00%82.00%84.00%86.00%88.00%90.00%92.00%94.00%96.00%98.00%ModelC-C5ModelB-LogisticRegressionModelA-NeuralNetworkNO.1ActualSeizurePotentialSeizureHealthyOverallAccuracy-TruePredictionCoincidenceMatrixofLogisticRegressionActualSeizurePotentialSeizureHealthyActualSeizurePotentialSeizureHealthy43387386320201033294TruePredictionofEitherTypeTruePredictionofNon-SeizureActivityTruePredictionofSeizureActivityRESULTANALYSIS&MODELVALIDATION3AspectsofValidationinT/FDiagnose97.16%97.00%97.12%99.18%79.23%95.12%96.99%95.72%96.73%75.00%80.00%85.00%90.00%95.00%100.00%SPECIFICITYSENSITIVITYACCURACYModelC-LogisticRegressionModelB-DiscriminantModelA-NeuralNetworkRESULTANALYSIS&MODELVALIDATIONBEST!GainChartHowgoodeachmodelpredictingresultsapproachtheidealcondition?StrengthsoftheModel•QuickandAccurate.•Modelsaredynamic,andconstantlylearnfromnewdata.•Costeffective.Targetedfollow-uptreatment•Savestimeswhichbenefitspatients.LimitationsoftheModel•Potentialcostsofmisclassifyingapatient.•Modelhastobeusedasasupportivetool.•Modelisonlyasgoodasdatausedtobuildit.•AbnormalEEGmaybecausedbyotherneurologicaldisorders.PotentialPuremedical/researchusePervasivedomesticuse(App/Website)ContinuousEEGrecordDataSetforresearchpurposeIndividualpatient’scustomizationPotential8.766.243.193.171.691.591.541.391.371.340510DeathsinmillionsTop10causesofdeathsgloballyIschaemicheartdiseaseStrokeChronicobstructivepulmonarydiseaseTracheaandbronchuscancerslungcancerDiabetesDementiasLowerrespiratoryinfectionsDiarrhoealdiseasesTuberculosisImplicationsforotherdiseasesSeizuredisordersHeadinjuryEncephalitisBraintumorencephalopathyMemoryproblemsSleepdisordersStroke(2rd)Dementia(7th)Thankyouforyourtime!AnyQuestions?