Different machine learning methods for tic-tac-toe prediction

Abstract

As a chess game, tic-tac-toe has a long history, and many chess players have a deep understanding of this chess game. With the advancement of science and technology, people can analyze daily affairs through the ability of computer deep learning. This paper hopes to use some machine learning models to analyze the tic-tac-toe chess game, so that the analysis results can improve the game players’ ability to play. This paper uses several different models for data analysis of the tic-tac-toe endings data-set, however, the purpose of doing this kind of research is to build an AI that allows players to play tic-tac-toe games and let game players gain richer game experience. For data analysis, this paper selects three analysis models that are more suitable for the outcome of tic-tac-toe game, which are Random Forest, Decision Tree, and SVM. After data analysis, this paper obtained the following results. The model accuracy of decision tree is 98.57%, random forest model accuracy is 95.61% and the SVM model accuracy is 98.33%