Research on Stock Price Prediction Based on Machine Learning Techniques

Hongyu Yao

2025

Abstract

Considering that financial stock markets are volatile and non-linear, accurately predicting stock closure price values is difficult. With the development of powerful machine learning methods and enhanced capacity for computation, predicting stock prices using machine learning methods is preferred because of its efficiency and effectiveness. In this project, Linear Regression (LR), Long Short-Term Memory (LSTM) and Gated Recurrent Unit (GRU) algorithms have been used to forecast closing price values of Tesla. The original financial data -- close price is regarded as the target variable, then open, high, low prices are used to calculate new features. To avoid multicollinearity issues, only volume, Relative Strength Index(RSI) and high-low ratio features are used as inputs for modelling part. Based on the standard strategic metrics, LR performs the best with the lowest RMSE 6.8703, the lowest MAE 4.0410, and the highest R-squared (R2) 0.9705. All metrics results suggest that LR has the most accurate results among all models. Furthermore, this article applies the residual plot and Quantile-Quantile plot to assess LR’s fit, in order to ensure its reliability and robustness.

Paper Citation

in Harvard Style

Yao H. (2025). Research on Stock Price Prediction Based on Machine Learning Techniques. In Proceedings of the 2nd International Conference on Data Science and Engineering - Volume 1: ICDSE; ISBN 978-989-758-765-8, SciTePress, pages 654-659. DOI: 10.5220/0013703600004670

in Bibtex Style

@conference{icdse25,
author={Hongyu Yao},
title={Research on Stock Price Prediction Based on Machine Learning Techniques},
booktitle={Proceedings of the 2nd International Conference on Data Science and Engineering - Volume 1: ICDSE},
year={2025},
pages={654-659},
publisher={SciTePress},
organization={INSTICC},
doi={10.5220/0013703600004670},
isbn={978-989-758-765-8},
}

in EndNote Style

TY - CONF

JO - Proceedings of the 2nd International Conference on Data Science and Engineering - Volume 1: ICDSE
TI - Research on Stock Price Prediction Based on Machine Learning Techniques
SN - 978-989-758-765-8
AU - Yao H.
PY - 2025
SP - 654
EP - 659
DO - 10.5220/0013703600004670
PB - SciTePress