Presenting a Smart Model for Distinguishing Patients With Thyroid Diseases From Healthy People by Combining Particle Swarm Optimization Algorithm and Artificial Neural Network

Thyroid diseases are common disorders worldwide. The timely diagnosis and control of this disease can prevent its progression and reduce associated complications. This study proposes a novel hybrid method by combining particle swarm optimization (PSO) algorithm and artificial neural network (ANN) for the timely detection of thyroid disorders.

This is an applied survey study, conducted in 2022. In this study, the target population consisted of the data of 400 patients referred to Imam Reza Hospital in Lar County, Iran from 2021 to 2022 which were collected by field study. Among them, 300 had thyroid disease and 100 were healthy. MATLAB software was used for implementing the proposed learning model and analyzing the results.

The regression coefficient of the proposed model in there modes of training, validation, and testing were 0.98, 0.97, and 0.95, respectively. The area under the ROC curve was 0.98, the error rate was 0.004, and the overall accuracy was 96%.

The proposed model can distinguish patients with thyroid disease from healthy individuals with acceptable accuracy and low errors. This model can be used as a useful tool in predicting thyroid diseases.