Detecting the variables affecting on Bitcoin price: Bayesian Model Averaging and Weighted Averaging Least Square approach

The purpose of this paper detecting the variables affecting on Bitcoin price using daily Time series data from 2015 to 2019 invoking two method of Bayesian model Averaging and Weighted-Average Least Square. The results of this study show that the price variables of cryptocurrencies with different creation mechanisms from Bitcoin and also the number of circulating cryptocurrencies with similar mechanism to Bitcoin and the volume of liquidity of US dollars affect the price of Bitcoin. On the other hand, the Forex market currency pairs, such as the dollar to Canadian dollar, the dollar to Australian dollar and the dollar to New Zealand dollar, which are less valuable than other major currency pairs in the Forex market, affect the price of Bitcoin. Also, the variables in the number of bitcoins, the number of cryptocurrencies in circulation with a different mechanism from bitcoin, the global price of gold and the number of searches for the word bitcoin in Google on its price have low coefficients. Overall, the results of the two methods of Bayesian averaging and Weighted Averaging Least Square are largely the same, and the use of the optimal pattern selection method confirms this.