Discussing the article: "Neural Networks in Trading: Memory Augmented Context-Aware Learning for Cryptocurrency Markets (Final Part)"

[MetaQuotes]

Check out the new article: Neural Networks in Trading: Memory Augmented Context-Aware Learning for Cryptocurrency Markets (Final Part).

The MacroHFT framework for high-frequency cryptocurrency trading uses context-aware reinforcement learning and memory to adapt to dynamic market conditions. At the end of this article, we will test the implemented approaches on real historical data to assess their effectiveness.

We have completed extensive work implementing our interpretation of the approaches proposed by the MacroHFT framework authors using MQL5. The next step is to evaluate the effectiveness of the implemented methods on real historical data.

It should be noted that the implementation presented here differs significantly from the original, including in the choice of technical indicators. This will inevitably affect the results, so any conclusions are preliminary and specific to these modifications.

For model training, we used EURUSD data from 2024 on the 1-minute timeframe (M1). The analyzed indicator parameters were left unchanged to focus on evaluating the algorithms and approaches themselves, without confounding effects from indicator settings. The procedure for collecting the training dataset and training the model was described above.

The trained model was tested on historical data from January 2025. The test results are presented below.

Author: Dmitriy Gizlyk