Comparative Analysis of Stock Market Trend Prediction Using Machine Learning and Deep Learning Algorithms on Continuous and Binary Data

Abstract

Financial market forecasting remains a complex challenge due to the volatile and unpredictable nature of stock price movements. This investigation presents a comprehensive evaluation of computational intelligence approaches for stock market trend analysis, focusing on reducing investment risks through advanced algorithmic predictions. Our study examines four distinct market sectors from the Tehran Stock Exchange: diversified financial services, petroleum industry, non-metallic mineral resources, and basic metallurgy. We implement and assess eleven predictive algorithms, including nine traditional machine learning approaches (Decision Tree, Random Forest, Adaptive Boosting, eXtreme Gradient Boosting, Support Vector Classifier, Naïve Bayes, K-Nearest Neighbors, Logistic Regression, and Artificial Neural Network) alongside two sophisticated deep learning methodologies (Recurrent Neural Network and Long Short-Term Memory networks). The analysis utilizes ten technical market indicators derived from a decade of historical trading data, processed through both continuous and binary data transformation approaches.