Prognozowanie kursu bitcoina z wykorzystaniem sztucznej sieci neuronowej
Forecasting the bitcoin rate using an artificial neural network

Summary/Abstract: Purpose – The aim of the article is to present the concept of an IT system that allows forecasting of the bitcoin cryptocurrency rate in relation to the euro. Design/methodology/approach – For the needs of pursuing such a formulated goal, an artificial neural network model – a multi-layer perceptron was developed. As part of the research, the input variables on which the BTC exchange rate was based were selected. Relevant data from daily quotations of exchange rates of selected currencies and metals were also obtained. The data was subjected to appropriate mathematical processing in order to adapt them for use during teaching, validation and testing of the artificial neural network. Originality/value – The results of the conducted experiments confirmed the high effectiveness of forecasting in one and two days perspective. High values of the regression coefficient (R) and small mean square error (MSE) indicate that the developed prediction system correctly predicts the rates of the analyzed cryptocurrency not only in relation to historical data, but also for current and future values.

• Details
• Contents