Discussing the article: "Overcoming The Limitation of Machine Learning (Part 7): Automatic Strategy Selection"

New comment
 

Check out the new article: Overcoming The Limitation of Machine Learning (Part 7): Automatic Strategy Selection.

This article demonstrates how to automatically identify potentially profitable trading strategies using MetaTrader 5. White-box solutions, powered by unsupervised matrix factorization, are faster to configure, more interpretable, and provide clear guidance on which strategies to retain. Black-box solutions, while more time-consuming, are better suited for complex market conditions that white-box approaches may not capture. Join us as we discuss how our trading strategies can help us carefully identify profitable strategies under any circumstance.

We live in an age of unprecedented interconnectivity—an informational revolution. But what happens when new ideas appear and spread faster than any trader can evaluate them? Given countless possible strategies, how can we automatically identify a shortlist we believe is worth testing? Can we discover potentially profitable configurations of strategies, without brute-forcing every possible combination?

This article proposes a framework to address these questions through two complementary approaches:

  1. White Box Solution: Use matrix factorization—specifically singular value decomposition (SVD)—on expected returns to identify strategy combinations positively influenced by current market conditions.
  2. Black Box Solution: Employ deep neural networks to dynamically select strategies based on observed market behavior.

Our solution relies on our ability to estimate the returns that would've been generated by following the trading strategies on hand. We then leverage our understanding of numerical computing to learn the expected revenue streams from our strategies. There is valuable insight to be gained from approximating the returns generated by any given strategy.

Author: Gamuchirai Zororo Ndawana

New comment