Poker-AI.org Poker AI and Botting Discussion Forum 2013-10-18T22:12:45+00:00 http://poker-ai.org/phpbb/feed.php?f=24&t=2615 2013-10-18T22:12:45+00:00 2013-10-18T22:12:45+00:00 http://poker-ai.org/phpbb/viewtopic.php?t=2615&p=5111#p5111 <![CDATA[Re: ANN Input Normalization]]> Statistics: Posted by cantina — Fri Oct 18, 2013 10:12 pm

]]> 2013-10-18T20:20:50+00:00 2013-10-18T20:20:50+00:00 http://poker-ai.org/phpbb/viewtopic.php?t=2615&p=5106#p5106 <![CDATA[ANN Input Normalization]]>
This question goes for those that are using Neural Networks (for determining actions or opponent modeling): How do you normalize the money/chips values (bankroll, amount to call, pot size, etc..) on a NoLimit SnG? ANNs needs those values to be normalized to the range (0..1) or (-1..1), depending on the activation function, but in a NLSNG, this values vary to much (because of the blinds increase).

I am planning to normalize it using the total chips in the tournament as the max. Ex: 10 players, 1000 initial bankroll, max will be 10000. At least for the current bankroll that should work, right? What do you think?

Thanks,
Corintio

Statistics: Posted by corintio — Fri Oct 18, 2013 8:20 pm

]]>