Analyzing the machine Learning methods to predict Bitcoin Pricing

Abstract

A cryptocurrency is a form of digital money that uses the blockchain technology and cryptography to protect the information about transactions and exchange made on the digital market. A cryptocurrency like Bitcoin consists of a big network that has many peers working on it and every peer has a record of the whole history that contains all the transactions that ever happened. Bitcoin is the most popular cryptocurrency. Bitcoin has attracted a lot of attention from individual and institutional investors. The purpose of this paper is to analyse the machine learning methods to predict Bitcoin pricing. Machine learning and its associated fields have made notable advances in recent years. Machine learning techniques is used in different areas of science particularly cryptocurrency price forecasting. Using this machine learning model, we can predict the price direction of Bitcoin. Machine learning methods have been demonstrated to be effective in predicting bitcoin prices. Few machine learning models to predict the Bitcoin price are Linear Regression, Logistic Regression, Bayesian Regression, Support Vendor Machine, Random Forest, Neural Network were discussed. Every method of machine learning has its own advantages and disadvantages, however from the literature analysis it is understood that the Artificial Neural Network and Support Vendor Machine have the highest effectiveness rate. Machine learning methods have higher prediction accuracy than parametric regression approaches.