I have continued my exploration on using Machine Learning algorithms for Finance specifically in Investing with a Master student who is working in an investment institution.

In the work done for the master’s project, significant improvements were obtained for the portfolio returns using deep reinforcement learning algorithm. An open-source deep reinforcement learning (DRL) framework is available here AI4Finance-Foundation/FinRL: FinRL: Financial Reinforcement Learning. 🔥 (github.com). The documentations are detailed out in this website Welcome to FinRL Library! — FinRL 0.3.1 documentation.

We have obtained an improvement of portfolio return for a period of one year using the DRL, after applying the algorithm with a higher return of 4.4% and 1.1% compared to the Modern Portfolio Theory (MTP), and equal weight portfolio. The stocks that were in the basket of the portfolio is twelve different sectors Syariah-compliance stocks within the KLSE market.

Most of the projects done in the program mostly focuses on short term – daily trading algorithms. However, this project was the first attempt here to evaluate machine learning algorithms for portfolio optimization in investing that is more crucial for fund managers.

There are also other resources focusing on this area, a course by Ashwin Rao, where you can obtain his book available here CME 241: Foundations of Reinforcement Learning with Applications in Finance (stanford.edu). Another interesting reference for AI/ML in Investing is a company that I have followed for some time – Euclidean Technology Value Investing — Euclidean Technologies ®

Signing off: 1/3/2023