JournalofMachineLearningResearch13(2012)3253-3295Submitted10/11,Revised9/12;Published11/12LinearFitted-QIterationwithMultipleRewardFunctionsDanielJ.LizotteDLIZOTTE@UWATERLOO.CADavidR.CheritonSchoolofComputerScienceUniversityofWaterlooWaterloo,ONN2L3G1,Can - PPT Presentation
LIZOTTEBOWLINGANDMURPHYHereaitrepresentstheactiontreatmentattimeandoitrepresentsmeasurementsmadeofpatientiafteractionait1andbeforeactionaitTherstobservationsoi1arebaselinemeasurementsmadebefor ID: 89312
LIZOTTE
BOWLINGANDMURPHYHere
aitrepresentstheaction(treatment)attime