Predictive analysis of the S&P 500 stock price index using Machine Learning
Journal of High School Science · 2024
DOI: 10.64336/001c.120714 ↗Kurz gefasst
Fachartikel in der Zeitschrift Journal of High School Science, 2024. Verfasst von Sihyun Lee und 1 weiteren Autor:innen. Der Volltext ist frei zugänglich (Open Access, CC BY-NC-SA 4.0).
Abstract (Originalsprache)
We attempted to predict the future trend of the S&P 500 (Standard and Poor’s 500) stock market index using machine learning algorithms. Unlike the fundamental analysis which is affected by intangible economic circumstances, we hypothesized that using only technical analysis, we could effectively predict future index movement based on historical trends if the targets were well defined within a period of the movement. This technical approach takes advantage of the indicators defined by mathematical or statistical measures, called the features. We chose price, volume change, day gain, 50- and 200-day Simple Moving Average (SMA), Relative Strength Index (RSI), and Moving Average Convergence Divergence (MACD), as the features. Initially, the S&P 500 index between Jan 19, 1993, and Feb 17, 2023 was represented with these features. After splitting the index into training and learning sets, the future trends were predicted with two supervised classification algorithms; decision tree and random forest, for the targets. We set the targets as next day, next week, and next month. We found that prediction accuracy increased as the defined prediction period increased. For example, the prediction accuracy for the next month was ~ 80%, as compared with ~ 50% for that of the next day. Prioritizing the indicators’ importance suggested that the SMA contributed the most to prediction accuracy. In addition, we investigated serval stock price trends of companies representing sustainable environment engineering, finance, and Information Technology in the commercial market using the same approach. These individual stocks trended comparable with the S&P 500 stock index analysis. Since SMA is a process to smooth out the trends over long transitions, it is less sensitive to time-related variation and capable of predicting long-term trends with minimal indefinite fluctuations. Thus, as the prediction horizon increased, the indicators that projected longer trends became more critical.
Zitieren
APA
Lee, S., & Sung, Z. (2024). Predictive analysis of the S&P 500 stock price index using Machine Learning. Journal of High School Science. https://doi.org/10.64336/001c.120714
Vancouver
Lee S, Sung Z. Predictive analysis of the S&P 500 stock price index using Machine Learning. Journal of High School Science. 2024. doi:10.64336/001c.120714
BibTeX
@article{Lee2024,
author = {Lee, Sihyun and Sung, Zuhawn},
title = {Predictive analysis of the S&P 500 stock price index using Machine Learning},
journal = {Journal of High School Science},
year = {2024},
doi = {10.64336/001c.120714}
}
EUVIMED