Comparative Analysis of Various Machine Learning Approaches for Bitcoin Price Prediction

Bitcoin is used worldwide for digital payment or simply for investment purposes. Bitcoin price prediction is an interesting research problem in current scenario. In this paper, we have studied the application of machine learning approach in predicting the future price of bitcoin. Many dynamic factors effect Bitcoin prices and accurate predictions form strong base for investment decisions. In this study, we have collected the live data corresponding to bitcoin from quindle.com containing 8 features. Then we have compared the prediction performance of 11 regression algorithms. It is found that Lasso regression with a combination of generalised linear regression outperformed others with an improvement of 9 % accuracy over other regression algorithms.