Abstract:

Currently, the importance of cryptocurrencies is increasing. While cryptocurrencies are used on virtual gaming platforms, they are now used as a payment tool by many institutions and organizations. The production is provided with the algorithm called blockchain (Block Chain) against security risks. The correct prediction of cryptocurrency prices is of great importance for investors and decision makers. In this study, the price values of the four commonly used cryptocurrencies (Bitcoin, Ethereum, Ripple, Litecoin) were estimated. The analysis was extended by identifying the fractures of each series using the multi-breaking test. It has been used both the classic time series models and three different types of artificial nerve network models to correctly predict the value of the captured virtual currency. Simply combination techniques have also been applied on the obtained predictions. Among these series that rational walk is dominated, it has been observed that better results are achieved than the rational walk model in Bitcoin virtual currency, with the volume of transactions and the knowledge.

Keywords: