Butterfly Optimized Feature Selection with Fuzzy C-Means Classifier for Thyroid Prediction

Abstract

The main task of thyroid hormones is controlling the metabolism rate of humans, the development of neurons, and the significant growth of reproductive activities. In medical science, thyroid disorder will lead to creating thyroiditis and thyroid cancer. The two main thyroid disorders are hyperthyroidism and hypothyroidism. Many research works focus on the prediction of thyroid disorder. To improve the accuracy in the classification of thyroid disorder this paper proposes optimization-based feature selection by using differential evolution with the Butterfly optimization algorithm (DE-BOA). For the classifier fuzzy C-means algorithm (FCM) is used. The proposed DEBOA-FCM is evaluated with parametric metric measures of sensitivity, specificity, and accuracy. In this work, the thyroid disease dataset collected from the machine learning University of California Irvine (UCI) database was used. The accuracy rate for the Differential Evolutionary algorithm got 0.884, the Butterfly optimization algorithm got 0.906, Fuzzy C-Means algorithm got 0.899 and DEBOA + Focused Concept Miner (FCM) proposed work 0.943.