Take the 2-minute tour ×
Mathematics Stack Exchange is a question and answer site for people studying math at any level and professionals in related fields. It's 100% free, no registration required.

I am attempting to write the multinomial Naive Bayes Classifier Decision rule as a linear rule. A document is viewed as a sequence $d = (w_1, w_2,\ldots,w_l)$ of $l$ words and is classified as $$h(d) = \operatorname{argmax}_{y\in\{+1,-1\}}P(Y=y)\prod_{i=1}^l P(W=w_i\mid Y=y)$$ Assuming we have estimates for all the appropriate probabilities and none of them are zero, how can we rewrite as $h(d) = \operatorname{sign}(\vec{v}\cdot\vec{x} + b)$? What is $\vec{v}$, $\vec{x}$, and $b$?

share|improve this question
add comment

1 Answer 1

In this generality, it could be almost anything. One possibility that comes to mind is

$$\operatorname{sign}\left(\sum_{y\in\{+1,-1\}}y\,P(Y=y)\prod_{i=1}^{l}P(W=w_i\mid Y=y)\right)\;,$$

share|improve this answer
add comment

Your Answer

 
discard

By posting your answer, you agree to the privacy policy and terms of service.

Not the answer you're looking for? Browse other questions tagged or ask your own question.