Contents I. Introduction
Machine learning (ML) has recently gained popularity in the medical field. One of its most important applications is the digital diagnosis of diseases. Machine learning is currently used for the early detection of many diseases such as cancer, diabetes, and thyroid due to its capacity to work with large amounts of data. However, there are several chal-lenges when implementing machine learning techniques in healthcare, notably the high dimensionality of clinical data, which increases the complexity of the classification task and reduces its efficiency. Another challenge is the imbalance in the class distribution in medical datasets, which leads to a decrease in generalization in machine learning algorithms. Furthermore, the machine learning method in healthcare must be as accurate as possible because a misclassification could lead to disease progression or even death. Several machine-learning models for thyroid disease prediction have been developed. Thyroid disease occurs when the thyroid gland releases abnormal amounts of hormones. The thyroid gland is essential in the body. It secretes hormones that regulate metabolism and growth in every cell, tissue, and organ. Hy-pothyroidism and hyperthyroidism are two common thyroid disorders. Hypothyroidism corresponds to insufficient thyroid hormone production, while hyperthyroidism corresponds to excessive production [1].
