<![CDATA[The increasing volatility and complexity of cryptocurrency markets have led to the growing application of machine learning (ML) techniques for accurate price prediction. This study presents a comparative analysis of eleven recent research papers on cryptocurrency forecasting using various ML and deep learning models, including Support Vector Machines (SVM), Random Forests (RF), Long Short-Term Memory (LSTM), Gated Recurrent Units (GRU), and ensemble methods. The findings highlight that deep learning models, particularly GRU and LSTM, often outperform traditional statistical models in capturing non-linear patterns and temporal dependencies. Moreover, feature diversity—such as on-chain data, market sentiment, and macroeconomic indicators—has been shown to significantly enhance predictive performance. However, many studies still lack comprehensive validation strategies and rely solely on historical price data, limiting generalizability. This review identifies key gaps in model benchmarking, feature integration, and evaluation consistency, providing a foundation for future research focused on hybrid models and interpretable AI for financial decision-making.]]>
Copyrights © 2024
