Predictive Analytics in Stock Market Forecasting: A Comparison of Decision Trees and Support Vector Machines

Abstract

Stock market prediction, as one of the critical challenges in the financial domain, requires the application of advanced machine learning algorithms. Decision Trees and Support Vector Machines (SVM) are two popular algorithms, each with its advantages and limitations in predicting market fluctuations. The aim of this paper is to compare the performance, accuracy, and efficiency of Decision Trees and Support Vector Machines in stock market forecasting. This study follows a descriptive-analytical review method, examining past research on the use of Decision Trees and Support Vector Machines for stock market prediction. Relevant studies from reputable academic sources were analyzed based on accuracy, speed, and the ability of each algorithm to process complex data. Decision Trees, due to their simplicity and high processing speed, perform well in quick and straightforward predictions, but they may struggle with overfitting when dealing with complex and noisy data. In contrast, Support Vector Machines, utilizing kernel and optimization techniques, demonstrate higher accuracy in identifying complex patterns and nonlinear data, though they require more complex parameter tuning and longer computational time. Both algorithms can be effective tools for stock market forecasting, depending on the type of data and market conditions. Decision Trees are better suited for quick and interpretable predictions, while Support Vector Machines outperform in handling complex and noisy datasets.