Main Article Content
This Bayesian network model was developed by analyzing the correlation between the cause of disease symptom variables and disease variables. The Bayesian network is a method that can depict causality between variables in a system. In this research, the Bayesian network was developed with a scoring based method and it was implemented using a hill-climbing algorithm with scoring BIC score function approach. There were 18 variables and 31 arcs representing the interconnection between symptom variable and respiratory tract disease. In the testing phase, the inference process using approximate inference was carried out and the accuracy was nearly 100% for all testing scenarios. The application of this method could result in a representative Bayesian network. Its resulted structure was affected so much by data condition, thus data cleaning was important to do before the training and testing phase.