Improvising healthcare decision making by employing ensemble technique

Abstract

Today, World is facing an elevated threat from a multitude of diseases that are growing at an astronomical rate every day. The numbers of diseases detected by medical institutions are increasing every year. Early prediction or detection of any disease can help people cure it to the fullest. After the initial cure, whether disease will create further health issues in future is also the area of investigation. Thus, predicting disease is a more important task to help clinicians to provide effective treatment for people. In this paper, we combine several classification approaches to improve the accuracy of the classifier. We propose an iterative ensemble approach that constructs a powerful classifier by mixing manifold low-performance classifiers with the intention that a powerful classifier with high precision can be obtained. The dataset used in this work maintains approx 50 attributes of diabetic patients. We examine whether after the initial recovery the patient has health issue in future or not.