Data mining with decision tree to evaluate the pattern on effectiveness of treatment for pulmonary tuberculosis: A clustering and classification techniques.

Scientific Research Journal; 2015; III; 43-48.

Abstract : Data mining is a process which helps in uncovering interesting data patterns in large volume of data. This procedure has become an ever more activity in all areas of medical science research especially in healthcare circumstances. Data mining has resulted in the innovation of useful hidden patterns from enormous databases. In this paper, a methodology is proposed for the programmed exposure and classification to evaluate the pattern on effectiveness of treatment for Pulmonary Tuberculosis (PTB) patients. Tuberculosis is a disease caused by mycobacterium which spreads through the air and hits low immune bodies easily. Our methodology is based on clustering and classification that classifies the success rate of Tuberculosis treatment based on the two broad classifications of the drug susceptibility testing (DST) namely, sensitivity to all drugs and resistance to any one drug. Age and weight are the main influencing factors for PTB patients, Two Step Clustering (TSC) is used to group data into different clusters and assign classes based on age and weight besides, The same procedure is being compared between with and without clusters of age and weight. Subsequently multiple different classification algorithms are trained on the result set to build the final classifier model based on decision tree along with K-fold cross validation method. The best obtained treatment effectiveness was 97.9% on a specified pattern from Classification and Regression Trees (CART). The proposed approach helps clinicians in their treatment planning procedures for different categories (through decision trees) to discover relationships which are currently hidden in the data.

Back to List of publications / Home