Bias Parameter in MLP in spark

Asked Aug 03 '16 at 06:10

Active Aug 03 '16 at 10:09

Viewed 223 times

I am using MLP of spark mlib for predictive analysis. I want to give more weightage to one of the output class and i am using Multilayer perceptron classifier for prediction. As given in the documentation here Spark Multilayer perceptron classifier.

       y(x)=fK(...f2(wT2f1(wT1x+b1)+b2)...+bK)

Is there any give by which i can give my own bias parameter (i.e. b1, b2 etc) in above given classifier? Reason to do so is that i want to reduce the false negative. Moreover, Is this the correct way to do so?

edited Aug 03 '16 at 10:09

Naresh

5,073
12
67
124

asked Aug 03 '16 at 06:10

deenbandhu

Your question isn't clear, please review it ! – eliasah Aug 03 '16 at 06:30
@eliasah let me know what confusion do you have in my question sorry for poor explanation of problem. – deenbandhu Aug 03 '16 at 06:46
I have edited the question kindly reply if you have any idea about this thanks in advance – deenbandhu Aug 03 '16 at 10:10

Bias Parameter in MLP in spark

0 Answers0