An empirical study and analysis of disease prediction risk assessment and data analytics using machine learning techniques in medical diagnostics

Abstract

Data mining plays a vital role as a tool offering numerous applications in healthcare industry by fetching knowledge and information on decision making. Some of the application of data mining in healthcare includes treatment effectiveness, drug discovery, healthcare management, decision support system, patients Electronic Health Record (EHR), improve service and treatment management. A decision support system can effectively integrate different types of data, knowledge, and models to support healthcare professionals and business owners in decision making. Generally decision support system varies with respect to the type of services such as data driven decision support, knowledge driven decision support, communication driven decision support and model based decision support. The proposed decision support framework acts as a supporting tool for doctors on disease classification, disease risk estimation and predictive services on risks of hospitalization to improve service and management. The proposed framework in the study focuses on three distinctive parts (i) Disease Classification (ii) Risk Assessment and (iii) Predictive analytics. newlineThe proposed framework averts the lapses in detecting the presence or absence of diseases through proposed Ensemble Learning Feature Selection and classification technique; risk assessment involves ruling out the features contributing to disease risk and their associations between different classes using novel decision tree method with probability function ; Predictive modelling aims to improve service performances through identifying significant predictor variables and thereby using their interaction effects for risk of hospitalization of chronic diseases predominantly heart disease risk patients. newline newline