Using CNN and LSTM to predict popular chess game openings from the first few moves of any chess game.
Chess engines are more powerful and accurate now than they’ve ever been However, they are often not very easily implemented, as high end computing power is needed As such, a method to use machine learning to mimic the output of a chess engine was devised This method uses machine learning techniques to identify a chess opening based on its first several moves and to predict the winner of the game accordingly
Dataset was taken from Lichess online database Data contains all stats, including player elo, player name, opening, win-loss, etc. Data was then reduced to just the game’s moves, the win-loss, and the opening name Since there were multiple hundred openings, only the top 10 most popular were used
We developed a 5 layer fully connected CNN architecture with RELU activation. Each layer consists of a convolutional layer, a maxpooling layer, a batch normalization layer, and a ReLu activation function The output is then fed into a fully connected network that uses batch normalization, ReLu, and Dropout
