根据贝氏定理为基础.ppt

《根据贝氏定理为基础.ppt》由会员上传分享，免费在线阅读，更多相关内容在行业资料-天天文库。

1、Chapter5StatisticalMethodsByJinn-YiYehPh.D.4/7/2009Outline5.1STATISTICALINFERENCE5.2ASSESSINGDIFFERENCESINDATASETS5.3BAYESIANINFERENCE5.4PREDICTIVEREGRESSION5.5ANALYSISOFVARIANCE5.6LOGISTICREGRESSION5.7LOG-LINEARMODELS5.8LINEARDISCRIMINANTANALYSIS5.1STATISTICALINFERENCEDesc

2、riptivestatisticsV.SStatisticalinferencePopulation,Sample,DatasetParameterV.SStatisticInferencemethods:estimation,andtestsofhypotheses5.1STATISTICALINFERENCE(cont.)Estimation:ThegoalistogaininformationfromadatasetTinordertoestimateoneormoreparameterswbelongingtothemodelofth

3、ereal-worldsystemf(X,w)5.1STATISTICALINFERENCE(cont.)statisticaltesting:todecidewhetherahypothesisconcerningthevalueofthepopulationcharacteristicshouldbeacceptedorrejectednullhypothesisV.Salternativehypothesis5.2ASSESSINGDIFFERENCESINDATASETScentraltendency1235.2ASSESSINGDI

4、FFERENCESINDATASETS(cont.)datadispersion125.2ASSESSINGDIFFERENCESINDATASETS(cont.)BoxplotInmanystatisticalsoftwaretools,apopularlyusedvisualizationtoolofdescriptivestatisticalmeasuresforcentraltendencyanddispersion5.3BAYESIANINFERENCENaïveBayesianClassificationProcess(Simpl

5、eBayesianClassified)根據貝氏定理為基礎，用以判斷未知類別的資料應該最接近哪一個類別監督式學習方式(需訓練資料)P(H/X):事後機率P(H):事前機率5.3BAYESIANINFERENCE(cont.)GivenanadditionaldatasampleX(itsclassisunknown),itispossibletopredicttheclassforXusingthehighestconditionalprobabilityP(Ci/X)P(X):constantforallclasses，onlythepro

6、ductP(X/Ci)·P(Ci)needstobemaximizedP(Ci):Ci/m(mistotalnumberoftrainingsamples)5.3BAYESIANINFERENCE-exampleTable5.1:TrainingdatasetforaclassificationusingNaïveBayesianClassifierSampleAttribute1A1Attribute2A2Attribute3A3ClassC112112001132122412125012162222710115.3BAYESIANINFE

7、RENCE–example(cont.)Goal：topredictclassificationofthenewsampleX={1,2,2,class=?}maximizetheproductP(X/Ci)·P(Ci)fori=1,2Step1:computepriorprobabilitiesP(Ci)5.3BAYESIANINFERENCE–example(cont.)Step2:computeconditionalprobabilitiesP(xt/Ci)foreveryattributevaluegiveninthenewsampl

8、eX={1,2,2,C=?}5.3BAYESIANINFERENCE–example(cont.)Step3:Undertheassumptionofconditi