Exploring the landscape of cryptocurrency forecasting research: a bibliometric perspective

Abstract

Purpose The evolution of the cryptocurrency landscape has been innovative, dynamic and adaptable. Using performance analysis and science mapping techniques, the study aims to conduct a bibliometric analysis to examine the landscape of the cryptocurrency domain, focusing on the forecasting aspect. Design/methodology/approach The study uses metadata from the Scopus database, ranging from 2015 to 2024, comprising 849 articles. They identified significant research constituents and five major thematic clusters. Findings The findings suggest that the research in the domain has yet to reach its full potential. The clusters involve structural shifts or turbulence in cryptocurrency markets, machine learning-based cryptocurrency price prediction, forecasting Bitcoin price and volatility, Bitcoin returns analysis, and cryptocurrency: A hedge and safe haven alternative. Further empirical analysis revealed that the artificial neural network and deep neural network outperformed the traditional statistical model, the autoregressive integrated moving average (ARIMA). Originality/value The study supplement these findings with significant future research directions, which will be beneficial for upcoming studies as the field has immense potential and countless areas worth exploring.