Medical Diagnosis

The goal of this assignment was to learn the parameters of a Bayesian Network representing causal relationships between eight diagnosis using health records with missing values. We implemented the Expectation Maximization algorithm from scratch to learn the conditional probabilities of the network.