Supporting Algorithmic Trading with Machine Learning: Progress in Backend Technology

Over the last decades, the development of artificial intelligence revolutionized the finance sector. Algorithmic trading can be more beneficial not only because of the speed and frequency but also by ruling out the influence of human emotions. Over the years many programming languages were used for machine learning research and system development. Python being easy to learn but still powerful has become popular within the scientific computing community, with most recent machine learning libraries being Python-based. In this paper, Python optimization is tested on the machine learning models frequently used in algorithmic trading. The results are compared between four different Python versions: Python 3.7, Python 3.8, Python 3.9, and Python 3.11.