Deep Learning Models Meet Financial Data Modalities
Kasymkhan Khubiev, Mikhail Semenov
TL;DR
The paper addresses forecasting and trading signals from limit order book data using deep learning, proposing an image-based, DL-driven framework that treats LOB as a distinct financial modality. It introduces embedding and sampling techniques, including min-max domain scaling and two input representations (merged vs. stacked LOB images), and evaluates four CNN-based architectures (with CNN2LSTM) in high-frequency trading contexts. Key findings show that stacked LOB inputs and longer historical intervals improve forecasting accuracy, with the price-delta target favoring short horizons and returns aiding longer horizons, and that one-shot mixed-asset training can markedly reduce error. The work demonstrates practical trading insights under optimistic execution assumptions and points to future directions like reinforcement learning to enhance robustness and deployment in real markets.
Abstract
Algorithmic trading relies on extracting meaningful signals from diverse financial data sources, including candlestick charts, order statistics on put and canceled orders, traded volume data, limit order books, and news flow. While deep learning has demonstrated remarkable success in processing unstructured data and has significantly advanced natural language processing, its application to structured financial data remains an ongoing challenge. This study investigates the integration of deep learning models with financial data modalities, aiming to enhance predictive performance in trading strategies and portfolio optimization. We present a novel approach to incorporating limit order book analysis into algorithmic trading by developing embedding techniques and treating sequential limit order book snapshots as distinct input channels in an image-based representation. Our methodology for processing limit order book data achieves state-of-the-art performance in high-frequency trading algorithms, underscoring the effectiveness of deep learning in financial applications.