POPA ReconRegularization and Bagging: First Techniques in HFT Modelling
Regularization and Bagging: First Techniques in HFT Modelling Standard trading models (OLS) overfit the massive amount of noise in high-frequency trading data, leading to highly unstable predictions and poor performance in live environments. The Solution: Traders combine regularization (which adds a penalty to mathematically constrain model size) and bagging (which averages predictions across multiple randomly sampled versions of the data) to significantly reduce prediction variance and stabilize trading performance https://research.qfex.com/p/regularization-and-bagging-first
The excerpt is quoted from a public Telegram channel, which owns its text.