Nonlinear Autoregressive (NAR) Network with Bayesian Regularization for Accurate Prediction of Bitcoin Price

LES ATTENTES DE CESAIRE D’APRES SES ŒUVRES POETIQUES ET THEATRALES
August 31, 2018
Integrating Indigenous and Western Education: Curricular Implications for African Schools
August 31, 2018

Nonlinear Autoregressive (NAR) Network with Bayesian Regularization for Accurate Prediction of Bitcoin Price

Abstract

Cryptocurrency is a digital or virtual currency using cryptography for security purposes. A cryptocurrency is difficult to counterfeit because of its unique security feature. The key advantages of cryptocurrencies include easier fund transfers between parties and the facilitation of these transactions through the use of public and private keys for security purposes. These fund transfers are performed with lower processing costs by avoiding costly bank transactions. The most popular cryptocurrency is Bitcoin, which uses the blockchain technology. The blockchain technology provides a data structure that stores all data in an online ledger that can be copied across all computers running Bitcoin software. Every new block must be verified by ledgers of each user, which makes it almost impossible to forge transaction history. This paper uses the nonlinear autoregressive network with a Bayesian regularization training algorithm to analyze Bitcoin fluctuations from historical data and predict future prices.

DOWNLOAD FULL ARTICLE