Bibliometric Analysis of Machine Learning Models for Stock Price Prediction

Abstract

Machine learning models are becoming more popular for predicting stock prices. Their advanced algorithms often perform better than traditional statistical methods. This study aims to carry out a bibliometric analysis using VosViewer. It focuses on the number, variety, and connections of studies related to stock price prediction using Support Vector Machines (SVM), Random Forests, and Deep Learning techniques. Boolean operators were used to search for and filter relevant articles from the Dimensions database. We conducted a manual review to select only high-quality studies, resulting in 55 peer-reviewed articles for analysis.

The analysis highlights important trends, influential studies, and collaborative networks in the field. This provides a clear overview of how machine learning models are developing for stock price prediction. There has been a notable rise in publications since 2013, with major contributions from institutions in the United States, the United Kingdom, India, and China. Emerging themes, especially deep learning applications, have been identified. However, there is still a lack of in-depth academic evaluation of these models.

This study stresses the need for more empirical research to address algorithmic biases and improve model reliability. The findings offer useful insights for researchers and practitioners, guiding future studies and helping to develop more accurate stock price prediction models.