Analyzing the Applications of Supervised and Reinforcement AI and Machine Learning Models in Cryptocurrency Price Prediction

Abstract

Cryptocurrency markets account to over $2 trillion in terms of total capital in 2022, i.e., almost
similar to market capitalization of Apple at the same time. Cryptocurrencies have been widely
established in financial markets with huge sum of trades and transactions taking place every day.
Like other fiscal systems, price prediction is a major challenge in crypto trading. Hence, “Artificial
Intelligence (AI)” has been widely used to predict cryptocurrency prices and has become a wellknown matter to study in cryptocurrency.
Unlike legacy financial models, machine learning (ML) models have shown great performance in
finance. They are supposed to be best to deal with the problem of price prediction in the volatile
and complex crypto market. A lot of studies have been conducted on machine learning for
predicting movement and price as well as portfolio management. However, these models and
approaches are in nascent stages. This study reviews existing research on reinforcement and
supervised learning models to predict crypto prices. It also highlights potential areas to improve
and research gaps to fill. Additionally, it focuses on potential research directions and challenges
which will be interesting in machine learning and AI communities in crypto market.