View Single Post
Old 04-08-2013, 06:06 AM
kafar kafar is offline
Junior Member
Join Date: Apr 2013
Posts: 6
Default Re: *ANSWER* Question 9

interestingly, i tried a 10k runs of N=100. the avg number of iterations to converge is close to 500.

In my lazy implementation, I always pick the first misclassified point to update the weight vector. Looking at the theory, I don't see how that matters. But since I got this question wrong, I guess a randomly chosen misclassified point would works better.
Reply With Quote