A trainable chess engine with an integrated GUI which uses linear regression to make decisions.
- The chess engine is trained using the games played by renowned players.
- The engine makes use of a feature computer which returns a vector of feature values given a board state.
- Upon receiving this vector, it is matched with the knowledge base.
- The ability to match board states is learned using linear regression.
- After matching, the move to be played is determined from the board state closest to the current state.
- Execute the python script chess_engine.py
- Log on to
http://localhost:5000
in your browser. - Select one of the following modes,
- Training mode: Train your chess engine.
- Testing mode: Play a game of chess against your chess engine. Theme of the chess pieces and style of board can be changed from the files testing_page.html and index.js.
Chess engine training in progress