| Summary: | This study aims to assess the knowledge flow within the research field and provide recommendations for further investigation. Specifically, this study conducts a thematic analysis of articles published in peer-reviewed journals between 2014 and 2022. Two primary themes emerge from the co-occurring keywords: (1) cryptocurrency behaviour and (2) cryptocurrency price prediction models. The findings reveal the use of various methods for predicting cryptocurrency prices, including econometric and statistical approaches, machine learning (ML), deep learning (DL), and hybrid models. The overarching objective of all these models is to achieve optimal results in addressing the various challenges associated with predicting cryptocurrency prices. However, it is important to note that no single model can effectively address all the behavioural nuances within cryptocurrency price prediction datasets. To bridge this gap, we recommend that future researchers explore the development of a hybrid model that combines a statistical model with deep learning. Such a hybrid model has the potential to accurately address the behavioural challenges encountered in cryptocurrency price prediction data series. © 2025, Semarak Ilmu Publishing. All rights reserved.
|
|---|