Multi-Horizon Echo State Network Prediction of Intraday Stock Returns

2025-04-28

AI Summary

To address the high model complexity, substantial training costs, and poor real-time performance in intraday stock return forecasting, this paper proposes a multi-step rolling prediction framework based on Echo State Networks (ESNs). By employing a fixed, randomly initialized reservoir, the approach circumvents the gradient-based optimization bottlenecks inherent in conventional RNNs, thereby preserving strong temporal modeling capability while significantly improving both training efficiency and inference speed. This work represents the first systematic application of ESNs to multi-horizon intraday return prediction, integrated with domain-informed high-frequency financial feature engineering. Empirical evaluation on benchmark high-frequency datasets demonstrates that the proposed method reduces prediction error by 12–19% relative to LSTM and XGBoost, achieves sub-5-ms inference latency per prediction, and cuts training time by over 80%, thus delivering a compelling balance of high accuracy, low computational overhead, and strong real-time responsiveness.

Technology Category

• Machine Learning: Time-Series/Data Streams
• Natural Language Processing: Learning & Optimization for NLP
• Cognitive Modeling & Cognitive Systems: Neural Spike Coding

Application Category

• Search and Retrieval-Augmented AI: Efficiency and scalability of Web search engines
• Web Mining and Content Analysis: Large pretrained models with web data
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

Stock return prediction is a problem that has received much attention in the finance literature. In recent years, sophisticated machine learning methods have been shown to perform significantly better than ''classical'' prediction techniques. One downside of these approaches is that they are often very expensive to implement, for both training and inference, because of their high complexity. We propose a return prediction framework for intraday returns at multiple horizons based on Echo State Network (ESN) models, wherein a large portion of parameters are drawn at random and never trained. We show that this approach enjoys the benefits of recurrent neural network expressivity, inherently efficient implementation, and strong forecasting performance.

Problem

• Predicting intraday stock returns at multiple horizons
• Reducing high complexity costs in machine learning methods
• Utilizing Echo State Networks for efficient forecasting

Innovation

• Uses Echo State Network for stock prediction
• Randomly initializes untrained parameters
• Combines RNN expressivity with efficiency