A study and design of decision support system for the prognosis of rheumatoid arthritis using optimized classification approaches

Abstract

Rheumatoid Arthritis (RA) is the chronic auto immune disease that affects the entire human body especially joints. It results in disruptions in bones and its functions. The symptoms of this disease are unclear, only few people are identified with the symptoms at the earlier stage of the disease. In some cases, it takes several years to diagnose RA. Some viral disease shows similar symptoms alike of RA which make difficulties in appropriate diagnosis of the disease by General Physicians. Due to the lack of awareness and inadequate prediction methodologies, there is some delay in treating RA at earlier stage. This may result in irreversible joint damage. From the analysis, it is known that there are few studies related to RA prediction using computational approaches. Existing RA predictor models need to improve prediction accuracy and generalization. Hence, there is a need for strong decision support system to identify RA as early as possible. Thus, in this research, a clinical Decision Support System (DSS) is designed based on patient s data with RA and Non RA group and expert s medical opinion. Here, Machine learning algorithms are used for designing DSS to predict RA. There are no readily available RA data sources in India especially at TamilNadu. Hence the RA dataset constructed through Questionnaire from the patient s profile who have visited the outpatient unit of Sakthi Rheumatology Centre, Coimbatore from Jan 2018 to Dec 2018. In this research work, 1000 patients were involved, and the dataset consists of 20 features including clinical, laboratory and physical investigation consisting of 750 RA and 250 Non-RA patients. In machine learning algorithms, if all the features are included for classification, it takes more time and reduces the system performance. Hence, there is a need for feature selection algorithms to find the significant features of RA and it reduces the prediction time and improves the classification accuracy. The main aim of this research work is to design an effective decision support system using efficient feature selection and classification approaches based on machine learning and soft computing algorithm newline