choose the best class if 2 class have same P (c|d), naive bayes
Posted
by
ryandi
on Stack Overflow
See other posts from Stack Overflow
or by ryandi
Published on 2014-06-06T08:48:48Z
Indexed on
2014/06/06
9:25 UTC
Read the original article
Hit count: 214
Hello I have some question about naive bayes classifier . In my project I have to classify a text into a class from 4 available class.
In naive bayes we have formula like
cmap=argmax.P(d|c).P(c)
I have standarize the amount of training document of each class, so I got a same P(c)
value for each class (0.25).
Here's my question: What if a testing document token doesn't have any token which belong to any of those 4 class(in document training)?
Resulted to all of the class have same value of P(d|c).P(c)
.
Which class should i pick?
What if the token exist, and 2 class or more have same value of P(d|c).P(c)
what should I do?
Thank you..
© Stack Overflow or respective owner