Emerging Trends in Volatility Forecasting Using Machine Learning: A Bibliometric Analysis

Abstract

Volatility forecasting remains a cornerstone of financial economics, offering critical insights into risk management, asset valuation, and investment strategy development. With the increasing complexity of financial markets, machine learning (ML) and deep learning (DL) techniques have significantly influenced the way volatility in financial time series is modeled and analyzed. This study presents a bibliometric analysis of academic publications from 2000 to 2025 that explore the use of ML and DL techniques within the context of volatility forecasting. The analysis is based on data extracted from the Web of Science database, a leading source of reliable and comprehensive scholarly literature in this field. The study analyzes the methodological development of models, such as Long Short-Term Memory (LSTM) networks, Support Vector Machines (SVM), and other ML-based models, which stand out for their capacity to model the complex and nonlinear dynamics of financial time series. In recent years, hybrid modeling strategies, based on the integration of traditional statistical methods with artificial intelligence algorithms, have emerged as prominent approaches in volatility forecasting. The study also highlights emerging trends, such as the use of transformer architectures and meta-learning strategies in high-frequency trading markets and cryptocurrency markets, where volatility is especially pronounced. Comparative analyses across different asset types and timeframes offer a comprehensive framework for understanding how these models perform under various financial conditions.The analysis conducted in this study critically examines the evolving paradigms of volatility forecasting, tracking methodological innovations and scientific developments. In doing so, it underscores the potential of ML and DL techniques to enhance forecasting accuracy, with an expectation that advancements in this field will guide future research directions. The insights derived from the findings not only contribute to the existing academic literature but also facilitate a more effective visualization of the structural dynamics of the literature, analyzed through the VOSviewer software.

Keywords

• Volatility forecasting
• Machine learning
• LTSM
• SVM
• Financial time series

References

1. Alpaslan, B. (2025). Emerging trends in volatility forecasting using machine learning: A bibliometric analysis. The Eurasia Proceedings of Science, Technology, Engineering and Mathematics (EPSTEM), 33, 80-95.