|
P(X,Y) P(Y|X)P(X) P(Y|X)P(X)
P(X|Y) = ----------- = ----------------- = ------------------------------------
P(Y) P(Y) P(Y|X)P(X) + P(Y|not X)P(not X)
X : 슬픔, psychological factor, 자극 Y : 눈물, physiological factor, 반응 |
P(눈물|슬픔)P(슬픔) P(눈물|슬픔)P(슬픔)
P(슬픔|눈물) = --------------------- = -------------------------------------------------
P(눈물) P(눈물|슬픔)P(슬픔) + P(눈물|not 슬픔)P(not 슬픔)
--PuzzletChung P(질병판정|질병)P(질병) P(질병판정|질병)P(질병)
P(질병|질병판정) = ------------------------- = ---------------------------------------------------------
P(질병판정) P(질병판정|질병)P(질병) + P(질병판정|질병아님)P(질병아님)
P(질병판정|질병)P(질병)
= -----------------------------------
P(질병판정|질병)P(질병) + P(오진)
만약 P(질병판정|질병)P(질병) (실제로 질병에 걸린 사람이 질병이 있다고 판정을 받을 확률. 위 문제에는 언급이 없음)을 100%라고 한다면,1 * 0.001 0.001 ------------------- = ------- = 약 0.0196 (1 * 0.001) + 0.05 0.051
0.95 * 0.001 0.00095 --------------------------- = ------- = 약 0.01866 0.95 * 0.001 + 0.05 * 0.999 0.0509
1 * 0.001 0.001 --------------------- = ----- = 1 1 * 0.001 + 0 * 0.999 0.001
P(지우다|tag) P(tag) 0.25 * 0.006446
P(tag|지우다) = ----------------------- = ----------------- = 0.02417
P(지우다) 0.0666666
P(지우다|tag) P(tag) P(지우다|tag) P(tag)
P(tag|지우다) = ----------------------- = -----------------------------------------------------
P(지우다) P(지우다|tag) P(tag) + P(지우다|no tag)P(no tag)


(다만, 군데군데 자연스럽지 않은 번역이 좀 보이는군요)
물론 어떤 이론이건 한계가 있을 수 밖에 없겠지만, Bayesian statistics 는 어떤 부분에서는 거의 어쩔 수 없는 귀결인 경우가 많지 않나 생각됩니다. --지상은
On the theoretical side, a unifying framework for all machine-learning methods also has emerged since the late 1980s. This is the Bayesian probabilistic framework for modelling and inference. |
But why should the Bayesian approach be so compelling? Why use the language of probability theory, as opposed to any other method? The surprising answer to this question is that it can be proved, in a strict mathematical sense, that this is the only consistent way of reasoning in the presence of uncertainty. |
Finally, one should be aware that there is a more general set of axioms for a more complete theory that encompasses Bayesian probability theory. These are axioms of descision or utility theory, where the focus on how to take optimal decisions in the presence of uncertainty. Not suprisingly, the simple axioms of decision theory lead one to construct and estimate Bayesian probabilities associated with the uncertain envirionment, and to maximize the corresponding expected utility. In fact, an even more general theory is game theory, where the uncertain environment includes other agents or players |