These examples use the LoopyBeliefPropagation function to determine the marginal probability of the disease hepatitis by observing its symptoms.
Hepatitis is an inflammation of the liver that can be caused by drugs, alcohol, or (most often) a virus. Its most common symptoms are:
- Jaundice (yellowing of the skin and whites of the eyes)
- Internal bleeding
- Loss of appetite
- Fatigue
- Fever
- Dark urine
- Stupor
- Nausea/vomiting
Given the presence or absence of a given symptom, the LoopyBeliefPropagation function determines the conditional or marginal probability of hepatitis.
The following figure shows the DAG that represents the relationship between hepatitis and its symptoms. The DAG represents each symptom by a conditional node and the disease by the dependent, unobserved node. These examples assume that each observed node variable is independent and binary.
Relationship between Hepatitis and Symptoms
