Application of the Method of Multivariate Multi-stage Forecasting Based on the LSTM Deep Learning Model for Bitcoin Price Time Series

Abstract

Forecasting data and research on cryptocurrency price forecasting methods are increasing in importance. So far, methods based on LSTM deep learning architecture have shown the best results in forecasting cryptocurrency prices. In order to improve the accuracy of forecasting data, this paper investigates the application of a multivariate multistep forecasting method based on the LSTM deep learning model for the bitcoin price time series and evaluates its effectiveness. The variants of multivariate multistep forecasting implementation based on deep learning LSTM are analyzed, and a direct approach for building multistep forecasts is chosen. Time series of bitcoin price and cumulative stability and drawdowns are used as input data. Based on our research, we found that short-term predictions were most accurate using models trained on trading data. However, for long-term forecasts, incorporating stability features slightly improved accuracy.