Week6-1Probabilities

最新推荐文章于 2020-12-01 21:54:43 发布

zypandora

最新推荐文章于 2020-12-01 21:54:43 发布

阅读量236

点赞数

分类专栏： NLP(Michigan)

本文链接：https://blog.csdn.net/zypandora/article/details/50312055

版权

45 篇文章 0 订阅

订阅专栏

Probabilistic reasoning

peech recognition
- recognize speech and wreck a nice beach
machine translation
- l’avocat general: the attorney general and the general avocado
probabilities make it possible to combine evidence from multiple sources in a systematic way.

P (w 1, w 2, . . ., w n) = P (w m ∣ w 1, . . ., w n - 1) . . . P (w 2 ∣ w 1) P (w 1)

$P(w_1, w_2, ..., w_n) = P(w_m \mid w_1, ..., w_{n-1}) ... P(w_2 \mid w_1) P(w_1)$

P (A \cap B) = P (A) P (B) P (A ∣ B) = P (A)

$P( A \cap B) = P(A)P(B)\\ P(A \mid B) = P(A)$

Adding constraints
- P(walk = yes | weather = nice, freetime = yes, …)
- more accurate
- more difficult to estimate
Removing constraints(back off)
- P(walk = yes | weather = nice, freetime = yes, …)
- P(walk = yes | weather = nice, freetime = yes)
- P(walk = yes | weather = nice)
- less accurate but more statistically feasibile

关注