Philippine E-Journals

AN OPTIMAL LINEAR DISCRIMINANT AND PREDICTION MODEL FOR CLASSIFYING TUBERCULOSIS PATIENTS USING LINEAR PROGRAMMING TECHNIQUE

Janice Lee | Dennis T. Beng Hui

Discipline: Industrial Engineering

 

Abstract:

This paper made use of an LP model to develop a linear discriminant function for classifying tuberculosis patients as either Active Patients (Class A) or Inactive Patients (Class B). Class A patients are those that are infected with tuberculosis while Class B patients are those who are not infected by the disease. A total of 494 actual TB cases from The Quezon Institute (QI) were gathered and used as training and validation samples for the LP model and the discriminant function. The trained linear discriminant function was able to generate 78% accuracy in classifying a patient as Class A given that the patient actually has TB compared to the current method of detecting TB which was estimated to be at 60%.