Ensemble Classifier System for Automatic Diagnosis of Thyroid Disease

Abstract

Thyroid gland secretes hormones that govern many of the functions in our body, such as the way the body uses energy, consumes oxygen and produces heat. Thyroid disorders typically occur when this gland releases too many or too few hormones. An overactive or underactive thyroid can lead to a wide range of health problems. Automatic diagnosis of Thyroid disease via proper interpretation of the thyroid data set is an important classification problem. Thyroid disease dataset which is taken from UCI machine learning database was used. The proposed method uses both Multilayer Perceptron (MLP) and Cascade-Forward Back Propagation Network (CFBN) as base classifiers for the proposed combined classifier systems. The combined classifier is based on varying the parameters related to both the design and training of neural network classifiers. The proposed method achieved accuracy value 96.23% for both combined MLP and combined CFBN classifiers. It has been observed that these results are one of the best results compared with results obtained from related previous studies and reported in the UCI web sites. The experimental results obtained show that the proposed combined classifier can be successfully used for diagnosing thyroid disease.