ERROR ASSESSMENT IN FORECASTING CRYPTOCURRENCIES TRANSACTION COUNTS
USING VARIANTS OFTHE GREY LOTKA-VOLTERRA DYNAMICAL SYSTEM
Abstract
The error assessment is made on the classical Grey Model (GM(1,1)) and
the variants of Grey Lotka-Volterra dynamical system namely the Grey
Lotka-Volterra Model (GLVM), the Fractional Grey Lotka-Volterra Model
(FGLVM) and the Variable-order Fractional Grey Lotka-Volterra Model
(VFGLVM) for modeling the transaction counts of three selected
cryptocurrencies in 2-and 3-dimensional framework. Bitcoin, Litecoin and
Ripple. The cryptocurrencies of interest are Bitcoin, Litecoin and
Ripple. The 2-dimensional models use Bitcoin and Litecoin transactions
from April, 28, 2013 to February, 10, 2018. The 3-dimensional model uses
transactions of Bitcoin, Litecoin and Ripple from August, 7, 2013 to
February, 10, 2018. The error sequence patterns and the the Mean
Absolute Percentage Error (MAPE) suggest a relatively higher accuracy of
the VFLVM in 2- and 3-dimensional study.