A Bayesian network represents an uncertain system as a directed acyclic graph of variables and their conditional dependencies. Each variable depends on a defined subset of related variables, making relationships explicit while keeping reasoning manageable. section 2.4.1
When evidence is observed, the network updates probabilities by combining prior beliefs with the evidence. The multiplication rule combines conditional probabilities across connected variables; the addition rule keeps probabilities distributed across alternative outcomes. section 2.4.1
Medical diagnosis
A disease may cause symptoms that influence diagnostic-test results. Observing a symptom or test result updates the probability that the disease is present. section 2.4.1

