基于类随机游戏数据的Supervised Learning：Ouril游戏走棋评估模型开发

Hey everyone, I'm new to machine learning and just wrapped up some online foundational courses—so I decided to dive straight into my first hands-on project: building an AI bot for the traditional Ouril board game. Let me break down what this project is all about:

Game Rules Recap #

• Both players start with 6 "houses", each stocked with 4 stones initially.
• Players take turns picking one of their own houses, then distributing all the stones from that house one by one into consecutive houses (following the game's standard turn direction).
• You can capture your opponent's stones during the distribution process when specific in-game conditions are met.
• The first player to capture 25 stones wins the match.

Project Objective #

My core goal here is to train an evaluation function that can predict how strong a potential move is based on the current game state. Once this function is reliable, the bot will use it to select the optimal move on every turn, maximizing its chance of winning against human players or other bots.

内容的提问来源于stack exchange，提问作者sra