AN INVESTIGATION OF MACHINE LEARNING TECHNIQUES FOR CHRONIC KIDNEY DISEASE DETECTION

Santosh Reddy Chada; Srinath Vallakirthy

Volume : VIII, Issue : I, January - 2019

AN INVESTIGATION OF MACHINE LEARNING TECHNIQUES FOR CHRONIC KIDNEY DISEASE DETECTION

Santosh Reddy Chada, Srinath Vallakirthy

Abstract :

Nowadays, there has been a recent interest in applying machine learning (ML) techniques to detect different chronic diseases. Chronic kidney disease (CKD) is one such emerging global health problem, and its early detection can be beneficial to take precautions and preventive measures. Hence, there is a requirement of a more sophisticated and specialised system for detecting the CKD in early stages. Recently, ML techniques remain as a focus of researchers for detecting chronic diseases by keeping their potential benefits like adaptability, flexibility and learning by example into consideration. This work aims to present an investigation of various ML techniques for providing effective detection of CKD. It compares these techniques for their performance on a benchmark dataset of CKD in terms most common performance metrics. The results indicate that MLP is the most accurate ML technique in comparison for detecting the CKD up to 99.75%. Whereas, IB1 is the fastest technique that takes minimum time for building the model from CKD dataset. The comparative analysis of these techniques helps to identify the best performing ML technique. The recognized technique can be considered as a candidate for developing an effective CKD detection system. The proposed work also helps the readers and fellow researchers to better understand the framework for applying ML techniques to detect the disease like CKD.