Определение
байесов метод
Theбайесов методisbasedontheBayesianprincipleandusestheknowledgeofprobabilityandstatisticstoclassifythesampledataset.Duetoitssolidmathematicalfoundation,themisjudgmentrateofBayesianclassificationalgorithmisverylow.Thecharacteristicofбайесов методistocombinethepriorprobabilityandposteriorprobability,whichavoidsthesubjectivebiasofusingonlythepriorprobability,andalsoavoidstheover-fittingphenomenonofusingthesampleinformationalone.TheBayesianclassificationalgorithmshowsahigheraccuracyratewhenthedatasetislarge,andthealgorithmitselfisrelativelysimple.
Наивен байесов алгоритъм
Наивен байесов алгоритъмisoneofthemostwidelyusedclassificationalgorithms.
Naiveбайесов методisbasedontheBayesianalgorithm,whichiscorrespondinglysimplified,thatis,itisassumedthattheattributesareconditionallyindependentofeachotherwhenthetargetvalueisgiven.Thatistosay,noattributevariablehasalargerproportiontothedecisionresult,andnoattributevariablehasasmallerproportiontothedecisionresult.AlthoughthissimplificationmethodreducestheclassificationeffectoftheBayesianclassificationalgorithmtoacertainextent,inactualapplicationscenarios,itgreatlysimplifiesthecomplexityoftheбайесов метод.
Принцип на алгоритъма
NaiveBayesClassification(NBC)isamethodbasedonBayes'theoremandassumingthatthefeatureconditionsareindependentofeachother,firstthroughthegiventrainingSet,taketheindependencebetweenfeaturewordsasthepremise,learnthejointprobabilitydistributionfrominputtooutput,andthenbasedonthelearnedmodel,inputtofindtheoutputthatmaximizestheposteriorprobability.
Thereisasampledataset,andthecharacteristicattributesetofthecorrespondingsampledatais.Theclassvariableis,thatis,canbedividedintocategories.Whereismutuallyindependentandrandom,thepriorprobabilityofis,andtheposteriorprobabilityofis,CanbeobtainedbythenaiveBayesalgorithm,theposteriorprobabilitycanbecalculatedfromthepriorprobability,theevidence,theclassconditionalprobability:/p>
NaiveBayesisbasedontheindependenceofeachfeature.Inthecaseofagivencategoryof,theaboveformulaItcanbefurtherexpressedasthefollowingformula:
Fromtheabovetwoformulas,theposteriorprobabilitycanbecalculatedas:
Sincethesizeofisfixed,whencomparingposteriorprobabilities,onlythenumeratoroftheaboveformulacanbecompared.Therefore,anaiveBayesiancalculationwithsampledatabelongingtothecategorycanbeobtained:
предимства и недостатъци
Предимства
TheNaiveBayesalgorithmassumesthattheattributesofthedatasetareindependentofeachother.Therefore,thelogicofthealgorithmisverysimpleandthealgorithmisrelativelystable.Whenthedatapresentsdifferentcharacteristics,theNaiveBayesalgorithmTheclassificationperformanceofYeshwillnotbemuchdifferent.Inotherwords,therobustnessofthenaiveBayesalgorithmisbetter,anditwillnotshowmuchdifferencefordifferenttypesofdatasets.Whentherelationshipbetweentheattributesofthedatasetisrelativelyindependent,thenaiveBayesclassificationalgorithmwillhavebetterresults.
Недостатъци
TheconditionofattributeindependenceisalsotheshortcomingofthenaiveBayesclassifier.Theindependenceoftheattributesofthedatasetisdifficulttosatisfyinmanycases,becausetheattributesofthedatasetareoftenrelatedtoeachother.Ifthiskindofproblemoccursintheclassificationprocess,theeffectoftheclassificationwillbegreatlyreduced.
Приложение
Текстова класификация
Classificationisabasicprobleminthefieldofdataanalysisandmachinelearning.Текстова класификацияhasbeenwidelyusedinmanyaspectssuchasnetworkinformationfiltering,informationretrievalandinformationrecommendation.Data-drivenclassifierlearninghasalwaysbeenahotspotinrecentyears,withmanymethods,suchasneuralnetworks,decisiontrees,supportvectormachines,andnaiveBayes.Comparedwithotherwell-designedandmorecomplexclassificationalgorithms,thenaiveBayesclassificationalgorithmisoneoftheclassifierswithbetterlearningefficiencyandclassificationeffect.TheintuitivetextclassificationalgorithmisalsothesimplestBayesianclassifier.Ithasgoodinterpretability.ThecharacteristicofthenaiveBayesalgorithmisthatitassumesthattheappearanceofallfeaturesareindependentofeachotherandeachfeatureisequallyimportant.Butinfactthisassumptiondoesnotholdintherealworld:firstly,theinevitableconnectionbetweentwoadjacentwordscannotbeindependent;secondly,foranarticle,someoftherepresentativewordsdetermineitstheme.Thereisnoneedtoreadtheentirearticleandlookatallthewords.Therefore,itisnecessarytoadoptasuitablemethodforfeatureselection,sothatthenaiveBayesclassifiercanachievehigherclassificationefficiency.
други
NaiveBayesalgorithmplaysamoreimportantroleintextrecognitionandimagerecognitiondirection.Anunknowntextorimagecanbeclassifiedaccordingtoitsexistingclassificationrules,andfinallythepurposeofclassificationisachieved.
TheNaiveBayesalgorithmiswidelyusedinreallife,suchastextclassification,spamclassification,creditevaluation,phishingwebsitedetectionandsoon.