LOB-based deep learning models for stock price trend prediction: a benchmark study

Matteo Prata, Giuseppe Masi, Leonardo Berti, Viviana Arrigoni, Andrea Coletta, Irene Cannistraci, Svitlana Vyetrenko, Paola Velardi, Novella Bartolini · 2024 · Artificial Intelligence Review

IntermediateHFTEmpiricalFree to read

Re-runs 15 published deep LOB models on common data: most fail to generalise beyond FI-2010. A reality check before trusting DeepLOB-style results.

Read the paper ↗Free version ↗

At a glance

Level: Intermediate: Comfortable with statistics, regressions and basic probability models
Horizon: HFT
Data needed: L2 depth (snapshots) Set what data you have
Themes: Machine learning Backtesting
Assets: Equities
Market: Global

Your notes

Report a problem with this entry

Step 6 of 6 in

Deep learning on the order book (with a reality check)

← Deep order flow imbalance: Extracting alpha at multiple horizons from the limit order book

Cite

@article{prata2024based,
  author = {Prata, Matteo and Masi, Giuseppe and Berti, Leonardo and Arrigoni, Viviana and Coletta, Andrea and Cannistraci, Irene and Vyetrenko, Svitlana and Velardi, Paola and Bartolini, Novella},
  title = {{LOB-based deep learning models for stock price trend prediction: a benchmark study}},
  journal = {Artificial Intelligence Review},
  year = {2024},
  doi = {10.1007/s10462-024-10715-4}
}

Prata, M., Masi, G., Berti, L., Arrigoni, V., Coletta, A., Cannistraci, I., Vyetrenko, S., Velardi, P., & Bartolini, N. (2024). LOB-based deep learning models for stock price trend prediction: a benchmark study. Artificial Intelligence Review. https://doi.org/10.1007/s10462-024-10715-4