A Performance Analysis of Machine Learning Algorithms in Stock Market Prediction, Compared to Traditional Indicators

This study evaluates and compares the predictive capabilities of traditional technical analysis indicators and machine learning (ML) algorithms within the dynamic NASDAQ market setting. Utilizing a comparative methodology, historical data spanning a decade and comprising four key variables are analyzed. The findings reveal the superior performance of ML algorithms, frequently achieving accuracy levels exceeding 80%. Specifically, the Support Vector Machine (SVM) stands out as a robust performer, affirming its efficacy in trend prediction. On the other hand, despite the high level of accuracy of specific technical indicators, their rigidity and incapability of adjusting hyperparameters highlight the need for optimization. Thus, this paper has great importance for research on finance since it explains the benefits of both methods and stresses the need to take under consideration the specifics of market and data when choosing. Furthermore, this research provides valuable guidance for future endeavors in financial analysis and investment decision-making. It delineates a trajectory for future research, advocating for a comprehensive integration of technical and algorithmic methodologies. By continually refining and innovating upon existing analytical frameworks, the financial industry can stay abreast of evolving market dynamics and maintain a competitive edge in forecasting and decision-making. In conclusion, by proving the superiority of ML algorithms, especially SVM, and underlining the drawbacks of traditional indicators, this paper shows its importance for investment strategies. Furthermore, it defines the direction of future research, which should be aimed at comprehensive implementation of technical and algorithmic methods of financial decisions.