Tuberculosis Analysis by Naive Bayesian Classification

Tuberculosis (TB) is a social disease withmedical aspects. By the increasing availability of biomedicaland health-care data with a wide range ofcharacteristics, computer-based medical system isplaying an increasingly relevant role in assisting bothdiagnosis and treatment. Base on the knowledgestored, the system will learn the patterns using NaiveBayesian classification and decides the category of TBby probabilities. It is based on the theorem of posteriorprobability. This system intends to develop a diagnosissystem of automatic classification method for TBdiagnosis based on the symptoms of the patients. Thesystem stores the knowledge of the medical expertsand the medical records of the previous case asTraining database. This system also considers themissing value by filling data completely, because itneeds the actual symptoms of the patient forincreasing accuracy to classify. This system can givethe category of TB and treatment for the patient whohas TB symptoms by using Naive Bayesianclassification method on the Training database. Theaccuracy of the system for that patient is shown byusing hold-out method on testing database