Enhancement of features and characteristics of rough set and their implementation for developing an efficient medical diagnosis system

Abstract

In recent decades, there have been phenomenal improvements in the patient s health care newlinesystem including disease prevention, diagnosis, treatments and prognosis management. Recent newlineCOVID-19 outbreak and growing health risk are stemming towards a new dimension of newlinehealth care, prevailing the physical and mental disabilities in people. With the motto Diagnosis newlineIs Better Than Disease , the system aims in protecting patient s health and associated newlinecost. Patient history,diagnostic information are crucial in clinical decision making and with newlinetechnological support of bio sensors, IoT devices and mobile clinical equipments, the system newlinecollects large data and often raise issues of data uncertainty, ambiguity and dimensionality. newlineRough set theory (RST) manages the modeling of categories and decision-making problems newlinethat include vague, inaccurate, imprecise or incomplete information. With these sensitive newlineissues in mind, the culmination of this thesis is focused on the use of RST and its feature newlineadvancements and to develop models in addressing health care issues. Disease ontology newlineaddresses a comprehensive analysis of disease(s) defined within bio-medical reposirotaries in satisfying the needs of community. CBRS approximation has been used for information/ newlineknowledge acquisition, and to discover knowledge in spreading of VBD (Malaria) diseaseand#894; newlineMalaria-Dataset of 10 backward and tribal districts of Odisha adapted for experimental usage. newlineHigh-dimensionality, feature selection/ reduction are major concerns in health care newlinebased machine learning (ML) solutionsand#894; strategies in this study modeled with an advanced newlinepre-processing diagnosis method with Tolerance RS to achieve better accuracy. ML models newlineusing fuzzy c-means and Tolerance RS were proposed to classify breast cancer imaging and newlinelung X-ray imaging segments (covid/ pneumonia detection). Diagnosis errors and diagnostic newlinebias are two major concerns in medical diagnosis.