Re: PLA: Is randomly selecting the update important?
Thanks, I'll have to check out the proof for myself, but trust you're right. I assume the book contains a proof, just haven't gotten my hands on copy yet (in transit).
As for performance between selecting 1st mismatch vs. random mismatch, I doublechecked by running my code both ways a few times and random choice took considerably fewer iterations. Hopefully the other poster mixed it up when typing...
