Cryptocurrencies Price Index Prediction Using Neural Networks on Bittrex Exchange

摘要

Cryptocurrencies have become fairly popular in the market since they were first introduced in the early 2000s. Cryptocurrencies are used primarily outside existing banking and governmental institutions and are exchanged over the Internet. Cryptocurrency exchanges allow customers to trade cryptocurrencies for other assets, such as conventional fiat money, or to trade between different digital currencies. This paper proposes a method to predict fluctuations in the prices of cryptocurrencies, which are increasingly used for online transactions worldwide. A Multi-Layer Perceptron (MLP)-based Non Linear Autoregressive with Exogenous Inputs (NARX) cryptocurrencies price forecasting model using the closing past prices together with volume. The model is evaluated based on price data collected from Bittrex Exchange, a US-based famous cryptocurrency exchange. Validation tests and Prediction test indicate that the proposed model is suitable for predicting prices on collected data.