PIGEON INSPIRED OPTIMIZATION WITH DEEP BELIEF NETWORK FOR THYROID DISEASE DIAGNOSIS AND CLASSIFICATION

In last decades, thyroid diseases become a serious illness, which affects the thyroid glands owing to the raised level of thyroid hormones or any infections in the thyroid organs. Diagnosis of thyroid can be considered as a classification problem, and it is difficult to solve it by classical by traditional parametric and nonparametric statistical methods. Presently, machine learning (ML) and deep learning (DL) models appeared as proper tools for the disease diagnosis process. In this view, this paper presents a new Pigeon Inspired Optimization (PIO) with DL based Deep Belief Network (DBN) model, called PIO-DBN for thyroid disease diagnosis and classification. In PIO-DBN model, the input medical data is initially preprocessed to improve the data quality. Then, DBN based classification process takes place on the prepocessed data. Since hyperparameter tuning process is essential to achieve effective training process of any DL model, PIO algorithm is applied to tune the parameters of DBN model which is inspired from the foraging behavior of pigeons. Extensive experimentations were carried out on benchmark thyroid dataset and the results are investigated under different aspects. The experimental results ensured the effective diagnostic performance of the PIO-DBN model with the with the maximum accuracy of 98.91% and 96.28% on the thyroid dataset 1 and 2 respectively.