Train Recurrent Neural Network to Predict Stock Prices Using Daily Return Rate

Kangrong Shi

2024

Abstract

In financial markets, where stock prices are extremely volatile, predicting their future movements has always been a major challenge for the financial and academic communities. This study aims to explore a novel method of stock market price prediction, that is, using daily returns as training data, to replace the traditional forecasting models that rely on closing prices. Traditional forecasting models often fail to fully capture the complex patterns and nonlinear relationships of stock price dynamics, resulting in limited forecasting accuracy. In order to overcome these limitations, this study uses three advanced models in deep learning: Recurrent Neural Network (RNN), Long Short-Term Memory Network (LSTM), and Gated Recurrent Unit (GRU) to improve the accuracy of prediction through daily return data. Three key metrics were used in this experiment: Root Mean Square Error (RMSE), Mean Absolute Error (MAE), and Mean Square Error (MSE) to evaluate the performance of the model. These indicators can comprehensively measure the deviation between the predicted value of the model and the actual stock price, thus providing an important reference for the optimization and selection of the model. Through rigorous testing and comparison of these models, it can be found that models that use daily returns as input data have significant advantages in terms of forecast accuracy. This finding suggests that daily returns can provide more granular time-series information, which can help models capture short-term fluctuations in stock prices and market dynamics, thereby improving the accuracy of forecasts.

Paper Citation

in Harvard Style

Shi K. (2024). Train Recurrent Neural Network to Predict Stock Prices Using Daily Return Rate. In Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML; ISBN 978-989-758-754-2, SciTePress, pages 488-493. DOI: 10.5220/0013526600004619

in Bibtex Style

@conference{daml24,
author={Kangrong Shi},
title={Train Recurrent Neural Network to Predict Stock Prices Using Daily Return Rate},
booktitle={Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML},
year={2024},
pages={488-493},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013526600004619},
isbn={978-989-758-754-2},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 2nd International Conference on Data Analysis and Machine Learning - Volume 1: DAML
TI - Train Recurrent Neural Network to Predict Stock Prices Using Daily Return Rate
SN - 978-989-758-754-2
AU - Shi K.
PY - 2024
SP - 488
EP - 493
DO - 10.5220/0013526600004619
PB - SciTePress