-
Notifications
You must be signed in to change notification settings - Fork 1
Home
Guannings edited this page Jun 25, 2026
·
3 revisions
A regime-aware quantitative trading system combining an XGBoost ensemble classifier, an SLSQP portfolio optimizer, and hierarchical PPO reinforcement-learning agents.
➡️ Browse the source code & README
The complete mathematical appendix lives here in the wiki. Each section starts from first principles, builds the intuition, walks through the derivation, and ends with a concrete numerical example.
| Section | Topic | Key Question Answered |
|---|---|---|
| 0 | Foundational Concepts | What is a loss function? What is gradient descent? |
| A | XGBoost Ensemble Classifier | How does the ML classifier detect market regimes? |
| B | Objective Function & SLSQP Solver | How does the optimizer pick portfolio weights? |
| C | Shannon Entropy | How does the system measure diversification? |
| D | Geometric Brownian Motion | How are future stock prices simulated? |
| E | Proximal Policy Optimization (PPO) | How does the RL agent learn without destroying itself? |