閱讀101 返回首頁    go 技術社區[雲棲]


Bayes Rule

Bayes Rule

apply bayes rule to the probability of h given D.

P(h/D)hypothesis of data
= ( P(D/h)P(h) ) / P(D)

Chain rule
h and D together. (同時發生)

P(h/D)P(D) = P(D/h)P(h)

P(h,D) = P(D/h)P(h) or P(h,D) =P(h/D)P(D)

P(D) is prior of data
P(h) is prior of particular hypothesis space
prior is called domain knowladge for us

最後更新:2017-10-31 23:33:42

  上一篇:go  Ensemble learning: Boosting
  下一篇:go  INTERSPEECH 2017係列 | 語音識別技術之自適應技術