Comparative Analysis of Machine Learning and Traditional Forecasting Methods in Bitcoin Price Prediction

Ying Shen, Zhengxin Qi, Fernando Martinez, Mohamed Rahouti, D. Frank Hsu, Diogo Oliveira

Abstract: In an era where digital currencies are becoming increasingly prevalent, accurately predicting the price fluctuations of cryptocurrencies like Bitcoin is crucial for investors and analysts alike. This paper delves into this challenge by comparing the efficacy of traditional time series forecasting methods against modern machine learning (ML) techniques. Utilizing a rich dataset that includes Bitcoin’s technical indicators and publicly accessible financial data, the study methodically assesses the predictive performance of each model. The results underscore the superior forecasting capabilities of ML algorithms, with the bestperforming model significantly outstripping traditional methods in accuracy. This advancement not only showcases the potential of ML in enhancing financial market predictions but also sets a new benchmark for future research in cryptocurrency price forecasting.