Bitcoin is the most popular and widely used digital currency. Therefore, the prediction of the Bitcoin price movement is of great importance for the financial markets. In addition to econometric models, data mining methods are used in Bitcoin price estimation. With the help of the tools and methods used in data mining, the data is modeled and converted into information to be utilized. K-Star algorithm is an example based approach which is used in many fields such as data mining, object identification and control systems. In this study, the effect levels of Macroeconomic variables on Bitcoin prices are analyzed by using K-Star Algorithm based on Lazy Learning which is one of the Machine Learning Method. The data set of the study includes 510 observational values ​​of dependent and independent variables between 3 January 2017 - 30 January 2019. 474 (93%) of these observations are used for modeling (training) and 36 (7%) are used for classification (test). The success rate of the model is 61,1% on whether Bitcoin prices will “increase” or “decrease” in the next period, while the correct classification success rate on an increase rate is 71,42% and the correct classification success rate on a decrease rate is 46,66% on Bitcoin prices. As a result, it is found that Machine Learning Technique shows a certain performance but the predictability of Bitcoin prices is still below the expectations.