One of the most common gland disorders is thyroid disorders. Timely diagnosis of this disease can prevent the consequences and improve treatment process of the patient. For this reason, its timely and accurate diagnosis is of great significance, and researchers are always attempting new methods to diagnose this disease more accurately. Machine learning and neural networks are among the most widely used methods that have provided promising results in this regard. In this thesis, with the aim of increasing the detection accuracy of 3 classes of hyperthyroidism, hypothyroidism and normal, the combination of perceptron neural network and crow search algorithm was deployed. In this work, after performing pre-processing such as normalization and processing missing data, as well as dimensionality reduction, in the classification step, the perceptron classifier optimized with the crow search algorithm was employed. In this research, using the crow search algorithm, the optimal value for the weight and bias parameters of the neural network was determined. Finally, by using the proposed approach, we were able to achieve an accuracy of 98.9%. based on the results, the presented model outperformed other state of the art methods in this field.
