인공지능 및 기계학습 개론 I

discrete probability distributionindependent eventsidentically distributed according to binomial distributionMaximum Likelihood Estimation \- maximize

Bayes formulation$P(\\theta)$: the part of the prior knowledgeNeed to represent the prior knowledge wellmultiply goes smoothdoes not complicate the fo

a function mapping an event to a probabilityexamplef(x): distributionx: eventPDF, CDFvery commonly observed distributioncontinuous numeric valuelong t

More experience -> more prior knowledge -> imporve experienceThe effort of producing a better approximate functionPerfect world \- no observation err

Initialize h to the most specific in Hempty set -> include x1 -> include x2 ... -> include xnMany hypotheses possible, and No way to find the converge

The most accessible algorithmUse only A1a, b, ?(null data)Use only A9t, f