基于移动神经网络的心血管疾病诊断与治疗系统(IJIEEB-V9-N6-3)

I.J.InformationEngineeringandElectronicBusiness,2017,6,19-26PublishedOnlineNovember2017inMECS()DOI:10.5815/ijieeb.2017.06.03Copyright©2017MECSI.J.InformationEngineeringandElectronicBusiness,2017,6,19-26AMobile-basedNeuro-fuzzySystemforDiagnosingandTreatingCardiovascularDiseasesFolasadeO.IsinkayeEkitiStateUniversity/DepartmentofComputerScience,Ado-Ekiti,NigeriaEmail:[e-mail]nic/DepartmentofComputerScience,Ilaro,NigeriaEkitiStateUniversity/DepartmentofComputerScience,Ado-Ekiti,NigeriaEmail:{[e-mail],[e-mail]}Received:04November2016;Accepted:16December2016;Published:08November2017Abstract—Inourpresentenvironment,heartdiseasesareveryrampartandtheydescribethevarioustypesofdiseasesthataffecttheheart.Theyaccountfortheleadingcauseofdeathword-wideespecially,inAfrica.Itisthereforeveryimportantforindividualstohaveadequateknowledgeabouttheirhearthealthinordertoavoidtheriskofdecreasedlifeexpectancy.Thehighmortalityrateofheart(cardiovascular)diseasesisattributedtotheunequalratioofpatientstoscarcityofmedicalexpertswhocanprovidemedicalcare,alsopatientsarenotalwayswarntowaitinglonghoursonqueueinthehospital,especiallyincasesofemergency.ThispaperdesignedandimplementedaMobileNeuro-fuzzySystemthatusesthecombinationoftheintelligencetechniqueofArtificialNeuralNetworks(ANN)andthehuman-likereasoningstyleofFuzzyLogictodiagnoseandsuggestpossibletreatmentsforcardiovasculardiseasesthroughinteractivitywithuser.ItemploysprogramslikeMySQL,PHP,JAVA(Android)andXML(AndroidStudio)whiletoolslikeXAMPP,PhpStormandAndroidO/Swereusedtointegratethesetechniquestogether.Thesystem,provedtobeofenormousadvantageindiagnosingheartdiseases,asitdiagnosesandlearnsabouteachuserpertime,toprovideadequateandappropriateresultsandalsomakesreliablepredictionstousers.IndexTerms—Heartdisease,Neuro-fuzzysystem,ArtificialNeuralnetwork,IntelligenceTechnique,Android.I.INTRODUCTIONCardiovasculardiseasesencompassthevariousdiseasesthataffecttheheartandtheyaretheleadingcauseofdeathword-wide[1]especially,inAfrica.TheAmericanHeartAssociation(AHA)estimatesthat17.3milliondeathcasesarerecordedperyear,anumberthatisexpectedtogrowtomorethan23.6millionby2030[2].Heartfailurewasreportedastheprimarydiagnosisforhospitalizationamongmedicalcarebeneficiaries[3]Therearedifferentkindsofcardiovasculardiseaseswhichincludecoronaryarterydiseases[4]suchasanginaandmyocardialinfarctionwhichispopularlycalledheartattack[5]Theothercommonheartdiseasesarestroke,hypertensiveheartdisease[6],rheumaticheartdisease,cardiomyopathy,atrialfibrillation,congenitalheartdisease,endocarditis,peripheralarterydiseaseandvenousthrombosis.Thehighmortalityrateofcardiovasculardiseasesisattributedtotheunequalratioofpatientstoscarcityofmedicalexpertswhocanprovidemedicalcare.Thismortalityratehasconstantlydrawntheattentionofresearchersanddifferentsoftcomputingtechniqueshavebeendeployedtoreducethisrateandtoserveverylargeamountofpatientsinlesstime[7].Presently,mostoftheseresearchesfocusonmodelingpartsofhumanbodyandrecognizingdiseasesfromdifferentscanssuchascardiograms,CATscans(ComputerizedAxialTomographyscan),ultrasonicscans,andothers[8].ThispaperdevelopedaMobileNeuro-fuzzysystemthatusesthecombinationoftheadaptiveintelligenceofArtificialNeuralNetworks(ANN)andFuzzyLogictodiagnoseandsuggestpossibletreatmentsforcardiovasculardiseases.Therestofthepaperisorganizedasfollows;SectionIIreviewsrelatedworkonArtificialNeuralNetworkandFuzzyLogic,SectionIIIintroducesthedesigndetailoftheproposedsystemwhileSectionIVdescribestheimplementationdetailsofthesystemandSectionVconcludesthepaper.II.RELATEDWORKSArtificialNeuralNetworks(ANN)[9]andFuzzylogic[10][11](popularlyknownasNeuro-Fuzzy)arecurrentlydrawingresearchattentionintheareaofmedicalscience.ApplicationofANNinmedicalscienceincludesmodelinganddiagnosingcardiovasculardisorders[1][12]classificationanddiagnosticpredictionofcancers[13],diagnosisofurologicaldysfunctionsdiabetesdiseasediagnosis[14].Theyareusedintheanalysisofmedicalimagesfromavarietyofimagingmodalities[15].20AMobile-basedNeuro-fuzzySystemforDiagnosingandTreatingCardiovascularDiseasesCopyright©2017MECSI.J.InformationEngineeringandElectronicBusiness,2017,6,19-26Neuralnetworkasaclassificationtechniqueindataminingcanbeappliedtoextractrulesfromdiseasesinmedicaldiagnosis[10][16][17].Fromtherules,analysisandpredictionofdiseasescouldbedoneeasily.DifferentvariationsofANNhavebeenemployedtohelpalleviatetheproblemofdiagnosis,treatmentandmedicaladviceprovisioninginmedicine[18][19][20][21].Forexample,[7]designedaFuzzyExpertSystemforheartdiseasediagnosis.Thesystemhasthirteeninputfieldsandoneoutputfield.Inputfieldsconsistofheartdiseasesymptoms.Theoutputfieldreferstothepresenceofheartdiseaseinthepatient.Itusesintegervaluedfrom0(nopresence)to4todistinguishlevelsofpresenceofthedisease.ThesystemcanbeusedasanalternativeforexistingsystemstodistinguishthepresenceorabsenceofheartdiseaseVazirani[22]proposedamodularneuralnetwork