Machine (supervised) learning and board game

by vitor morev   Last Updated May 15, 2019 16:19 PM - source

Let's assume there's a board game like chess or go (in other words, perfect information game). We don't know how to write a good evaluation function for this type of board games, but we have a lot of played games with results. How can we use scikit-learn to predict a game result of a given position?

I tried classical minimax with alpha-beta pruning and greedy evaluation fuction, but it didn't work good when strategy (not only tactics) was required, so I decided to get an evaluation function by machine learning.



Related Questions



What *is* an Artificial Neural Network?

Updated August 16, 2018 01:19 AM