Nowadays, thyroid disease is common and spread worldwide. Various studies state that about forty two million people are experiencing thyroid effects in India. Finding of wellbeing circumstances is an enthusiastic and complicated task in the medical field. Increasing level in Thyroid Stimulating Hormone (TSH) or some contamination in thyroid glands causes thyroid disease. The classification of thyroid disease is made on the basis of the amount of hormones secreted. In case of hyperthyroidism, more hormones are secreted and in hypothyroidism, the secreted hormones will be less than the required hormones. In this paper, the Artificial Neural Network algorithms are compared and found the best predictive model of thyroid disease. The dataset is retrieved from UCI machine learning database and compared with the ANN algorithms. ANN is used as it is a non-linear data driven, adaptive and very powerful tool for predicting purposes. Here as an attempt, the prediction of thyroid disease is done using Levenberg-Marquardt (LM), Scaled Conjugate Gradient (SCG) and Bayesian Regularization (BR) algorithms of ANN. Performances yielded by ANN-LM, ANN-SCG and ANN-BR are compared and the best outcome is acquired from the multilayered perceptrons which is trained using Bayesian Regularization algorithm.
Volume 12 | 04-Special Issue
Pages: 636-646
DOI: 10.5373/JARDCS/V12SP4/20201529