Artificial Intelligence Techniques for Stock and Market Index Prediction: A PRISMA-Based Systematic Review (2018-2024)

Abstract

Background and Objective: Financial asset prices are shaped by numerous endogenous and exogenous factors, including macroeconomic conditions, political events and market sentiment. With advances in artificial intelligence and the availability of large-scale data, novel techniques have emerged for financial forecasting. This study aimed to systematically review recent methodologies for predicting stock and market index prices.

Materials and Methods: A Systematic Literature Review (SLR) was conducted to identify, screen and analyze relevant scientific publications. Searches were performed in the Scopus and IEEE Xplore databases, yielding 32 eligible studies published primarily since 2018, comprising five review articles and 27 original research papers.

Results: The analysis reveals that hybrid modeling approaches are the most frequently adopted techniques in recent studies. Forecasting is predominantly focused on short-term horizons and a diverse set of input features-including technical, fundamental and sentiment-based indicators-is employed across models.

Conclusion: Recent research trends emphasize hybrid and short-term forecasting models for financial assets, reflecting the growing integration of artificial intelligence with financial data analytics. These findings provide a structured overview of current practices and highlight directions for future research in financial market prediction.