RL4CAD: Personalized Decision Making for Coronary Artery Disease Treatment using Offline Reinforcement Learning

该地铁跑酷强化学习项目将原游戏封装到一个自定义的 Gymnasium 环境中。该环境通过mss进行实时屏幕捕获，使用 EasyOCR 进行数字识别来提取里程、金币和游戏结束信号，并通过pyautogui模拟鼠标滑动来控制角色。

End-to-end machine learning and reinforcement learning project for optimizing loan approval policies in a financial setting. Combines supervised deep learning (MLP) and offline reinforcement learning (CQL) to maximize profit and minimize risk using LendingClub Loan Data.

A comprehensive project exploring the transition from Supervised Deep Learning to Offline Reinforcement Learning for optimizing loan approval policies.