This project explores Bayesian network structure learning using mutual information and Minimum Description Length (MDL) scoring. It implements an algorithm to construct probabilistic graphical models based on real-world data and evaluate their efficiency in representing dependencies between variables.