Análisis comparativo de redes neuronales profundas para la predicción del precio de mercado de bitcoin

Abstract

One of the most challenging problems of recent years is the financial study of cryptocurrencies in order to know their future behavior. The rise of artificial intelligence is providing us with tools that allow us to address the issue with greater precision. In particular, deep learning allows us to generate neural networks that learn from large volumes of data, thus identifying patterns imperceptible to humans. In this paper, we focus on the study of Bitcoin, the most valuable and popular cryptocurrency on the market today. Three types of deep neural networks and a comparative model have been implemented when forecasting its price. The results of the comparative analysis conclude that temporal patterns turn out to be more representative than spatial patterns when estimating the behavior of Bitcoin’s future closing price.