Algoritmic Crypto Trading Utilizing Rule-Based and Machine Learning Approaches

Algorithmic trading is a popular approach for professional traders to increase their profitability and beat the markets. According to Yahoo Finance, algorithmic trading accounts for more than half of the overall US stock market trading and a large yearly growth is forecasted within the next decade. Similarly, algorithmic trading is growing in the crypto trading as well. Thanks to the easiness of reaching market data with new exchange APIs, algorithmic trading is much easier to be implemented. In addition to the easy reach of data, the Python language has multiple libraries, which allow even beginner level programmers to be able to implement algorithmic trading. In this thesis, I will research whether algorithmic trading is profitable by using machine learning and rule based strategies. Within the practical part, both rule-based and machine learning techniques, such as logistic regression and decision tree, will be implemented in python. The real-time data is taken from Binance Application Programming Interface (API) for implementing the trading strategy, using Python libraries, such as Numpy and Pandas for data handling and sklearn for machine learning. The strategies will be tried on different crypto coins at various timeframes, varying from 1 minute data to 1 day data frames.