Анализ и прогнозирование курсовой стоимости биткоина методом SSA

Abstract

In the paper the existing mathematical models of the financial market are reviewed. Nevertheless, the majority of the research published has the essential drawbacks such as the theoretical character of the papers, the amount of the observations required, the inadequate performance in the neighborhood of bifurcation points, and the absence of a computer model capable of real-time prediction. In the present paper we apply the SSA method for analyzing the structure and predicting the behavior of the bitcoin rate time series. The results obtained are of higher accuracy compared to the ones obtained by using the ARIMA and ARFIMA-FIGARCH models for predicting, even in the critical for these models cases.