Discussing the article: "Neural Networks in Trading: The Adaptive Graph Diffusion Model (Conclusion)"

[MetaQuotes]

Check out the new article: Neural Networks in Trading: The Adaptive Graph Diffusion Model (Conclusion).

In this article, we conclude our work on building the SAGDFN framework using MQL5, summarizing the development process and presenting the results of its practical testing. Let's combine the modules we've already implemented into a single system, highlight the strengths of this approach, point out its weaknesses, and discuss possible ways to improve it.

Training the model is like a well-planned expedition: before setting out into the open seas of the real market, we practiced thoroughly in the calm harbor of history. This first, offline phase was based on data from the EURUSD currency pair on the H1 timeframe for the entire year of 2024 — a period marked by significant contrasts. Here we found everything from the calm, almost mirror-smooth waters of sideways markets to the turbulent storms of sharp trend moves and unexpected gusts of news-driven volatility. This wide variety of market scenarios enabled the model to develop robust navigation and learn to recognize both common and rare price movement patterns, without losing its bearings even under challenging conditions.

Once this preparation was complete, it was time to leave the training dock and put the ship to the test on the currents of the real market. The second stage — online fine-tuning — was conducted in the real-world conditions of the MetaTrader 5 strategy tester. Here, data arrived sequentially, one candle after another, and the model learned not only to analyze streaming data but also to remain stable amid whirlpools of noise, on the shifting shoals of low liquidity, and during unexpected news squalls. This stage served as precision fine-tuning: it did not disrupt the framework that had already been established, but helped polish it for real-world conditions, increasing adaptability and reducing the risk of overfitting.

The final test turned out to be a real trial by fire. We took the data from January 2025 — completely new data, untouched by previous experiments — and loaded all the previously derived parameters without making a single change. This was a crucial point: no fine-tuning, no additional adjustments — just a pure test that reflected the model's true ability to generalize.

The test results are shown below.

Author: Dmitriy Gizlyk