Bitcoin price volitality prediction with machine learning

As a digital currency, Bitcoin has gained significant prominence in financial markets. However, the high volatility issue, a prominent characteristic of Bitcoin, is an intriguing topic for investors and academics. It can lead to sudden and large-scale changes in the price of Bitcoin, affecting investors' decision-making processes and making market analysis more challenging. This study aims to predict the price of Bitcoin by identifying the factors that affect the price movements of Bitcoin. Two different categories of data sources were used in this study. The first category consists of data related to the cryptocurrency system, including 350 variables encompassing Bitcoin and other cryptocurrencies. The second category includes data related to global indices, encompassing 46 variables. These data were obtained from popular financial data sources such as Tradingview and Yahoo Finance. The data relating to the cryptocurrency system were collected under headings such as network information, cryptocurrency exchange transactions, address data, price data, and social media data. The data related to global indices were classified under headings such as bond indices, currencies, volatility indices, and stock indices. Within the scope of the research, machine learning algorithms were employed to predict the price of Bitcoin using data encompassing the 90 days between October 2022 and December 2022. These data were divided into an 85% training set and a 15% test set for analysis. Following the training phase, predictions on the test data were visualized and presented. The obtained results have demonstrated that the model exhibited a generally successful performance with an R² score of 0.909, indicating its predictions were notably close to actual values. These findings substantiate the utility of machine learning algorithms as an effective tool for Bitcoin price prediction.