Systematic Literature Review: Advanced Artificial Intelligence Techniques for Forecasting Stock Prices

Abstract

Artificial intelligence-based models have recently secured a legitimate foothold in the financial markets, serving as powerful analytical tools capable of significantly reducing the inherent uncertainties of investment activities and assisting investors in identifying stocks with the highest profitability potential. A key advantage for users lies in the ability to ground their strategies and decisions not on subjective reasoning, but on objectively derived quantitative data. Financial time series forecasting stands as one of the most successful computational applications in finance, owing to both the diversity of its areas of application and its operational effectiveness. A considerable body of Machine Learning (ML) research has been dedicated to this task, resulting in a substantial volume of literature and numerous literature reviews conducted over the years. More recently, Deep Learning (DL) models have emerged as even more advanced alternatives, consistently outperforming traditional ML approaches. While the use of DL is rapidly expanding in financial contexts, systematic reviews focusing exclusively on this area remain scarce. Hence, the present review aims to fill this gap. This work provides a comprehensive and systematic review of the current literature related to the application of artificial intelligence in financial forecasting. The analysis is further segmented by investment domains such as stock markets, foreign exchange, and commodities and by the specific Deep Learning models applied in each area. Additionally, it includes a critical discussion of persistent challenges and future directions, offering valuable insights for scholars pursuing research in this emerging field.