TESTING METHODS AND MODELS TO FORECAST CRYPTOCURRENCIES EXCHANGE RATE
TESTING METHODS AND MODELS TO FORECAST CRYPTOCURRENCIES EXCHANGE RATE

Summary/Abstract: The course of cryptocurrencies forms by various factors which makes it difficult to apply fundamental methods for their forecasting. For these reasons technical analysis and various statistical models are used for short-term forex and financial market forecasting. In this study we test three models: the classical autoregression model (AR), the Box-Jenkins ARIMA, and the predictively modified model Frequency Analysis of the Volatility and Trend with movable calculation (FAVT-M). The five cryptocurrencies with the largest market capitalization as of July 10, 2019 are subject to test forecasting. The AR and ARIMA results report compromise confidence within the first 5 - 6 days, after which they show significant deviations from the actual course achieved. FAVT-M generates immediate signals for the reversal of the short-term trend, but at this stage they are not clear enough for its reliable independent application in forecasting cryptocurrencies.

• Details
• Contents