Evaluating Time-Series Models for Bitcoin Forecasting: From Exponential Smoothing to Deep Learning

Résumé fourni par la source

The emergence of cryptocurrencies has brought about a significant transformation in the financial industry. Beyond their investment potential, these assets have attracted portfolio investors due to their possible role in hedging strategies and diversifying traditional financial instruments. In this study, we employ three forecasting techniques, Autoregressive Integrated Moving Average (ARIMA), Facebook Prophet (Fb-Prophet), and Bidirectional Long Short-Term Memory (Bi-LSTM) to predict Bitcoin prices using historical market data from January 2012 to September 2020. Among the models, the Bi-LSTM achieved the lowest Root Mean Squared Error (RMSE) of 1.271 and the lowest Mean Absolute Error (MAE) of 2.314. Accurate Bitcoin price predictions enable investors to make informed decisions on when to buy, hold, or sell, thereby enhancing returns and reducing potential losses by identifying market trends and avoiding high-risk periods. For traders, especially in scalping or day trading, leveraging short-term price fluctuations can further improve profitability.

Ce résumé expose les affirmations des auteurs. BNTIC ne l’interprète pas comme une validation indépendante des résultats.