Application of Machine Learning in the Prediction of Hypothyreoidism

Hypothyroidism is a common condition that implies a reduced level of thyroid hormones. If this condition is not diagnosed in time, it can be fatal. The diagnosis of hypothyroidism is primarily based on clinical signs and symptoms, as well as measuring the concentration of thyroid-stimulating hormone (TSH). The database contained a total of 2000 samples. 1000 samples were healthy people and 1000 samples were sick people. Patients older than 18 years, of both sexes, were considered. The algorithm with the best performance on the test data was selected as the final model for predicting hypothyroidism. In this work, based on the used parameters and a database of 2000 patients, an expert model was developed whose training accuracy was 96%.