Hybridization of a Novel Rheumatoid Arthritisprediction Modelusing Classifier and Optimizer Approach

Abstract

Rheumatoid arthritis (RA) is an autoimmune illness causes systematic and chronic effects. It reduces physical functionality and causes articular and fatigue damages. Disease prediction is an application where Machine Learning tools pretend to give successful outcomes. Machine learning (ML) approaches help to predict RA at an earlier stage and reduce the severity caused due to RA. This work concentrates on two phases known as feature selection and classification with ML approaches. As Meta-heuristic optimizationaims to give a more optimal solution, it is integrated with ML based classifier model to give superior outcomes. This work considers hybrid Selfish optimization with Elman classifier model (HSO-EC) algorithm for feature selection and the chosen features are fed to the classifier model. In RA disease prediction, various individual classifier models try to give nominal outcomes; but fail in terms of prediction accuracy and precision. To overcome the drawbacks, this work considers a hybrid classifier model. The provided dataset is posed to handle classification problems with various numerical features. Thus, more influencing features are selected using (HSO-EC) algorithm and fed as an input to the classifier model. Data validation is performed with and performance metrics like F-measure, Recall, and Accuracy, Precision, ROC, sensitivity, and specificity are evaluated. The metrics are evaluated with the existing individual classifier model and the proposed hybrid classifier gives better accuracy for predicting RA.