Abstract:

Predictive techniques and models are extremely important for individuals and organizations in the right decision-making and investment phase. The accuracy of the forecast ensures that successful decisions are made and allows investors to a maximum benefit. In this study, the aim is to estimate the prices of the most commonly used Bitcoin from the types of cryptocurrencies using the artificial nervous networks method. As the input currencies; Dow-Jones, S&P500, Nasdaq100, Eurostoxx Indices, Swiss Frang, British Sterling, Euro, Gold, and Silver Investment Instruments were taken.   Daily closing prices data were used between 2013-2018. In the study, a re-nutrition artificial nerve network model was used. The model was tested by the January 2019 forecast and the model's forecast accuracy R2 was achieved with 99% success.

Keywords: