Hello,
is there anyway to change the cost of misclassification in the Mahout naive
bayes algorithm?  I have a binary classifier with a dataset consisting 98%
of Class A documents and 2% Class B docs. Naturally, to help with the
imbalance, I want to tell the classifier that the cost of misclassifying B
as A is greater than misclassifying A as B.

Best,
Harish Kumar

Reply via email to