What is it about?
Financial time series is hard to predict due to the low signal-to-noise ratio and the non-stationarity. We investigate the idea inspired from image classification in trading and the techniques from self-supervised learning in computer vision to financial time series to reduce the noise exposure and hence generate correct labels.
Featured Image
Photo by m. on Unsplash
Why is it important?
Our results show that these denoised labels improve the performances of the downstream learning algorithm, for both small and large datasets, while preserving the market trends. The findings also suggest that with our proposed techniques, self-supervised learning constitutes a powerful framework for generating “better” financial labels that are useful for studying the underlying patterns of the market.
Perspectives
We hope this research can provide a new and challenging direction to traders to design trading algorithms that perform well on denoised prices and can be backtested in various conditions.
Yanqing Ma
King's College London
Read the Original
This page is a summary of: Denoised Labels for Financial Time Series Data via Self-Supervised Learning, October 2022, ACM (Association for Computing Machinery),
DOI: 10.1145/3533271.3561687.
You can read the full text:
Contributors
The following have contributed to this page
