A Review on Artificial Intelligence-Driven Stock Market Forecasting: Advances, Challenges, and Future Directions

Ramsundar G, Radha D

Abstract

This review aims at reviewing the emerging trend associated with the use of artificial intelligence (AI) to support financial forecasting in the stock market. Traditional methods for financial prediction such as technical analysis and time-series modeling have several flaws in terms of the inability to accurately estimate financial indicators in a nonlinear, highly volatile market environment. On the contrary, the use of machine learning-based approaches is characterized by higher efficiency and scalability. Several popular approaches such as deep learning models, reinforcement learning (adaptive trading strategy development), natural language processing for sentiment analysis, and machine learning-based algorithms can be used in stock market forecasts. Advantages and disadvantages of these techniques are critically analyzed, and the emphasis is placed on ways to overcome the limitations associated with traditional forecasting techniques and increase predictive performance. In specific, the quality of data, different information sources, data preprocessing procedures, as well as key metrics are considered for evaluating models and their performances. Specific cases will be discussed in detail, and current limitations, including interpretability, dependency on data, and model opacity, will be outlined. Finally, future trends for research in this area will be highlighted.

Source: semanticscholar · PDF

Read the AI summary and key takeaways for traders on WOBR Quant Research.