Discussing the article: "Neural Networks in Trading: Anomaly Detection in the Frequency Domain (Final Part)"

New comment
 

Check out the new article: Neural Networks in Trading: Anomaly Detection in the Frequency Domain (Final Part).

We continue to work on implementing the CATCH framework, which combines the Fourier transform and frequency patching mechanisms, ensuring accurate detection of market anomalies. In this article, we complete the implementation of our own vision of the proposed approaches and test the new models on real historical data.

For training, a dataset was constructed using random episodes generated in the MetaTrader 5 Strategy Tester. The dataset is based on historical EURUSD M1 data covering the entire year of 2024.

Model evaluation was performed on historical data from January to March 2025. All experimental parameters were kept unchanged to ensure objectivity of the results and to enable an unbiased assessment of strategy performance. This setup guarantees that the model does not merely memorize the training dataset but instead demonstrates its ability to adapt to new market conditions.

The testing results are presented below.

Author: Dmitriy Gizlyk

New comment