Modified Full Bayesian Networks Classifiers for Medical Diagnosis

This paper is present a modified Bayesian Network (BN) for Full Bayesian Classifier (FBC) for used it to diagnoses the condition of a patient from some of symptoms and medical history. The conditions are the heart diseases and the nervous diseases. Modified for FBC is depended on common structure known as naïve Bayes. Determining the network structure is D-separated by the variable. Each variable has CPT and each disease (table) has Probability. By modified the equation for find CPT for each variable and Probability for each disease (table) in Modified-FBC (M-FBC) structure, the system for diagnosis is designed. The experimental resulted show that the successful ratio of heart diseases (93%) and nervous system diseases (98%) approximately.