A Literature Review on Diagnosing Thyroid Disease Through Artificial Neural Network Techniques

Abstract

In this study, researcher majorly focuses on usefulness and suitability of artificial neural network for prediction of thyroid disease and respective role of neural network in medical diagnosis by past recorded data is identified from a comprehensive literature review. Wherein, research contributions from 2013–2017 are reviewed. It is found that different various architecture of Artificial Neural Network (ANN), Back Propagation Network (BPN), Hybrid Back-Propagation Neural Network (BPNN), Data Mining Methods (DMM), Decision Support System (DSS), Ranked Improved F-score Ordering (RIFO), Auto Associative Neural Network (AANN), Multi-Layer Perception (MLP), Multivariate Bayesian Prediction Method (MBPM), Machine Learning Methods (MLM) and Radical Basis Function Neural Network (RBFNN) are found to be proper and appropriately suitable. In recent years, these architectures and methods are also found useful for prediction of so many other diseases. The discussions of these architectures and their suitability, appropriateness for thyroid disease prediction is presented through this review article and it would reveal the various methods and its benefits of diagnosing thyroid diseases in a most appropriate and essential manner based on the nature and cause it affects on human body. This review paper will certainly help various researchers to go through various research insights on diagnosing thyroid diseases on a single point of contact and it help them to focus and choose the appropriate method for diagnosing thyroid diseases.