A hybrid multiclass classifier based on artificial immune algorithm and support vector machine

Developing an effective medical diagnosis system for many diseases, such as thyroid gland disease, to assist physicians in hospitals has become a high priority for many researchers and clinical centers. In fact, existing medical diagnostic techniques often have to diagnose the risk of misdiagnosis. The purpose of this paper is to develop an efficient classifier to improve medical diagnosis performance of thyroid gland disease. In this work, the medical dataset of thyroid gland disease that represent multiclass classification problem was selected from the University of California Irvine Machine Learning Repository. The proposed approach combined support vector machines with an artificial immune system as the diagnostic classifier, which is called the AIS-based machine learning classifier. The diagnosis results were identified, and the accuracies of the classification rate were evaluated. The classification results demonstrated that the proposed approach can give considerable improvements over those reported in previous studies.