Computing Approach to Detect the Factors and their Impact in Prevalence of Thyroid

Abstract

Thyroid condition is often referred to as thyroid disorder when it makes more or less hormone secretion. Correlation analysis was performed among attributes to find the strongly correlated attributes, the entire model was trained using pre-existing classifiers to detect the prevalence of thyroid. The highly correlated attributes are also identified as the main risk factors for the occurrence of thyroid. Pre-existing classifiers were implemented as part of the training phase, which achieved an accuracy of 95% in order to improve the accuracy and minimize the error rate, a novel approach called ThyroNet was devised. The novel method has outperformed most of the pre-existing models with an accuracy of 97% and with a mere error rate of 0.0489. The study also highlights the factors influencing thyroid, its impact and identifies strategies for maintaining a healthy lifestyle.