Abstract
As the popularity of blockchain technology and investor confidence in Bitcoin (BTC) increased in recent years, many individuals started making BTC and other cryptocurrency investments, in expectation of high returns. However, as recent market movements have shown, the lack of regulation and oversight makes it difficult to guard against high volatility and potentially significant losses in this sector. In this study, we propose a data-driven Tree Augmented Naïve (TAN) Bayes methodology that can be used for identifying the most important factors (as well as their conditional, interdependent relationships) influencing BTC price movements. As the model is parsimonious without sacrificing accuracy, sensitivity, and specificity—as evident from the average accuracy value—the proposed methodology can be used in practice for making short-term investment decisions.