Hybrid Vector Auto Regression and Neural Network Model for Order Flow Imbalance Prediction in High Frequency Trading

TL;DR

This paper presents a hybrid VAR and neural network model that improves the prediction of Order Flow Imbalance in high frequency trading, aiding better market understanding and strategic decision-making.

Contribution

It introduces a novel hybrid model combining linear and non-linear methods for more accurate OFI prediction in high frequency trading.

Findings

Hybrid model outperforms standalone VAR and FNN models.

Significant improvement in predictive accuracy on real trading data.

Provides insights into buy and sell trading pressures.

Abstract

In high frequency trading, accurate prediction of Order Flow Imbalance (OFI) is crucial for understanding market dynamics and maintaining liquidity. This paper introduces a hybrid predictive model that combines Vector Auto Regression (VAR) with a simple feedforward neural network (FNN) to forecast OFI and assess trading intensity. The VAR component captures linear dependencies, while residuals are fed into the FNN to model non-linear patterns, enabling a comprehensive approach to OFI prediction. Additionally, the model calculates the intensity on the Buy or Sell side, providing insights into which side holds greater trading pressure. These insights facilitate the development of trading strategies by identifying periods of high buy or sell intensity. Using both synthetic and real trading data from Binance, we demonstrate that the hybrid model offers significant improvements in predictive accuracy and enhances strategic decision-making based on OFI dynamics. Furthermore, we compare the hybrid models performance with standalone FNN and VAR models, showing that the hybrid approach achieves superior forecasting accuracy across both synthetic and real datasets, making it the most effective model for OFI prediction in high frequency trading.