Short-Term Stock Price Direction Prediction Using Machine Learning Models

DOI

10.2991/978-94-6239-801-6_59How to use a DOI?

Keywords

Stock direction prediction; machine learning; return-based features; random forest; financial forecasting

Abstract

This study looks at a simple question in stock prediction. Can basic machine learning models give useful signals about whether a stock return will go up or down in the next period? The experiment does not try to predict the exact stock price. It uses relative returns, because returns are easier to compare across stocks and across time than absolute prices. The dataset contains five large technology stocks, and the features are built from lagged returns, moving averages and recent volatility. Four classifiers are tested under the same time-based training and testing setup, including logistic regression, random forest, gradient boosting and K-nearest neighbours. Random forest gives the best test accuracy in this experiment, reaching 0.630. Gradient boosting reaches 0.580, K-nearest neighbours reach 0.550 and logistic regression reaches 0.540. These results are slightly better than random guessing, but the improvement is still limited. The main message is that simple return-based features can provide some information, while short-term stock direction prediction remains difficult and would need richer data in future work.

Copyright

© 2026 The Author(s)

Open Access

Open Access This chapter is licensed under the terms of the Creative Commons Attribution-NonCommercial 4.0 International License (http://creativecommons.org/licenses/by-nc/4.0/), which permits any noncommercial use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons license and indicate if changes were made.

Download article (PDF)

TY  - CONF
AU  - Hang Sun
PY  - 2026
DA  - 2026/10/23
TI  - Short-Term Stock Price Direction Prediction Using Machine Learning Models
BT  - Proceedings of the 2026 2nd International Conference on Financial Innovation and Marketing Management (FIMM 2026)
PB  - Atlantis Press
SP  - 549
EP  - 555
SN  - 2352-5428
UR  - https://doi.org/10.2991/978-94-6239-801-6_59
DO  - 10.2991/978-94-6239-801-6_59
ID  - Sun2026
ER  -