Forecasting Ethereum Price by Tuned Long Short-Term Memory Model

Abstract

Cryptocurrencies have established a firm position in the economic world in the past decade, with thousands of distinctive currencies available for electronic payments. The majority of cryptocurrencies, however, experience extremely volatile price perturbations, drastically affecting investors and traders. To address this problem, this paper proposes long short-term memory approach tuned by salp swarm metaheuristics. This hybrid model has been validated on a benchmark financial dataset, and the outcomes have been compared to other cutting-edge methods. The results suggest that the proposed method outperformed the competitors, showing significant potential in time-series prediction tasks.