Tags
- PythonScikit-learnConsoleAI
About
A command-line Tic-Tac-Toe game that uses machine learning to determine the best move. The game was created as part of a project to practice machine learning with Scikit-learn and compare different algorithms. The player can choose between one of three algorithms: K-nearest neighbors, linear regression, and multi-layer perceptron. These algorithms are trained on a dataset of 6,500+ games. Each dataset entry contains the board state and the best moves to make. Once the player selects an algorithm, they can play against the computer and see how well the algorithm performs. KNN tends to perform the worst while MLP is nearly unbeatable.
Images
