A neural network informer in algorithmic investment strategies on high-frequency bitcoin data
To register for this seminar with Eventbrite please click here
Abstract:
The article investigates the usage of the Informer architecture for building automated trading strategies for high-frequency Bitcoin data. Three strategies using the Informer model with different loss functions: Root Mean Squared Error (RMSE), Generalized Mean Absolute Directional Loss (GMADL), and Quantile loss, are proposed and evaluated against the Buy and Hold benchmark and two benchmark strategies based on technical indicators. The evaluation is conducted using data of various frequencies: 5-min, 15-min, and 30-min intervals, over the 6 different periods. Although the Informer-based model with Quantile loss did not outperform the benchmark, two other models achieved better results. The performance of the model using RMSE loss worsens when used with higher frequency data, while the model that uses the novel GMADL loss function benefits from higher frequency data. When trained on 5-min intervals, it beats all the other strategies on most of the testing periods. The primary contribution of this study is the application and assessment of the RMSE, GMADL, and Quantile loss functions with the Informer model to forecast future returns, subsequently using these forecasts to develop automated trading strategies. The research provides evidence that employing an Informer model trained with the GMADL loss function can result in superior trading outcomes compared to the Buy&Hold approach.
Bio:
Robert Ślepaczuk is an Associate Professor at the University of Warsaw, where he heads the Department of Quantitative Finance and Machine Learning and the Quantitative Finance Research Group (QFRG). His research focuses on financial forecasting, algorithmic investment strategies, and applications of machine learning in financial markets. He has published in journals such as Expert Systems with Applications, Journal of Big Data, and Economic Modelling. He has more than 20 years of experience in algorithmic trading and quantitative investment strategies, combining academic research with practical industry experience. Previously, he served as Investment Director of the quantitative fund management division at Union Investment TFI. He is actively involved in bridging academic research with real-world trading applications and supervises projects in quantitative finance and AI-driven investment strategies.
Homepage: https://www.wne.uw.edu.pl/en/faculty/structure/departments/quantitative-finance