Probability Classification Model
内容来自于CS229,浙江大学机器学习听课笔记以及百度百科,blog补充
基本问题:假设有状态
w
1
,
w
2
w_1,w_2
w1,w2,样本
x
∈
{
w
1
,
w
2
}
x \in\{w_1,w_2\}
x∈{w1,w2},求
P
(
w
1
∣
x
)
,
P
(
w
2
∣
x
)
P(w_1|x),P(w_2|x)
P(w1∣x),P(w2∣x)
分类问题:
若
P
(
w
1
∣
x
)
>
P
(
w
2
∣
x
)
P(w_1|x)>P(w_2|x)
P(w1∣x)>P(w2∣x),则
x
∈
w
1
x\in w_1
x∈w1
若
P
(
w
1
∣
x
)
<
P
(
w
2
∣
x
)
P(w_1|x)<P(w_2|x)
P(w1∣x)<P(w2∣x),则
x
∈
w
2
x\in w_2
x∈w2
即:若
P
(
x
∣
w
1
)
P
(
w
1
)
>
P
(
x
∣
w
2
)
P
(
w
2
)
P(x|w_1)P(w_1)>P(x|w_2)P(w_2)
P(x∣w1)P(w1)>P(x∣w2)P(w2),则
x
∈
w
1
x\in w_1
x∈w1
若
P
(
x
∣
w
1
)
P
(
w
1
)
<
P
(
x
∣
w
2
)
P
(
w
2
)
P(x|w_1)P(w_1)<P(x|w_2)P(w_2)
P(x∣w1)P(w1)<P(x∣w2)P(w2),则
x
∈
w
2
x\in w_2
x∈w2
P
(
x
∣
w
)
P(x|w)
P(x∣w) – Conditional probability of x given w
P
(
w
∣
x
)
P(w|x)
P(w∣x) – Posterior probability of x given w
P
(
w
)
P(w)
P(w) – Prior probability of w
WARNING: Prior probability can be dicisive in certain circumstances.