Project for the course: Bio-Inspired Artificial Intelligence - Prof. Giovanni Iacca - University of Trento
Full Report: Project_LucaClaus.pdf
Code: Project.ipynb
This project explores the use of Genetic Programming (GP), a subset of Evolutionary Algorithms (EA), to develop a novel, profitable trading strategy for Bitcoin. As a highly volatile asset, Bitcoin provides a unique opportunity to evaluate the robustness of algorithmic trading strategies. The GP algorithm evolves mathematical expressions that model optimal trading behaviors by combining a range of technical indicators, including the Relative Strength Index (RSI), Moving Average Convergence Divergence (MACD), Bollinger Bands, and others. The study involves training the GP model on historical Bitcoin price data to maximize profitability. Backtesting on distinct datasets from different time periods ensures the strategy’s generalizability and adaptability to varying market conditions. Results are analyzed in terms of return of profit and consistency, providing insights into the viability of GP-based trading systems in dynamic financial markets.