Machine Learning Techniques for Stock Market Prediction: Neural Networks and Soft Computing Approaches

This paper provides an updated overview of recent literature on stock market prediction using machine learning methods. Neural network-based models, particularly those focused on predicting currency exchange rates, have become increasingly popular among researchers. A range of neural network architectures, including Artificial Neural Networks (ANN), Functional Link Artificial Neural Networks (FLANN), Radial Basis Function Networks (RBFN), Wavelet Neural Networks (WNN), and Psi Sigma Networks, are frequently integrated with other learning techniques such as Least Mean Squares (LMS), Genetic Algorithms (GA), and Particle Swarm Optimization (PSO) to improve both accuracy and efficiency. Given the inherent uncertainty in currency exchange models, soft computing approaches have proven effective in capturing the complex nonlinear dynamics of stock market data, resulting in enhanced forecasting outcomes.