Prediction of Thyroid Disease Using Random Forest Classifier and Hill Climbing Algorithm

Applying machine learning methods in thyroid diagnosis has significant importance in medical diagnostics. Thyroid diseases are common worldwide, and early diagnosis shows a crucial part of effective treatment and management. The main objective our study is to present a data-driven method that can accurately forecast thyroid disease in affected individuals. In this research, a combination of machine learning approaches such as Support Vector Classifier (SVC), K-Nearest Neighbors Classifier (KNNC), and Random Forest Classifier (RFC) using the Hill Climbing Algorithm (HCA), has been used to predict the thyroid disease, and the results demonstrate their effectiveness. The evaluation metrics in this paper include accuracy, precision, F1-score, and Matthews Correlation Coefficient (MCC), which compare the three proposed methods and show that the Random Forest technique with the hill climbing algorithm performs best, with values of 88.31, 89.43, 92.3, and 97 percent, respectively. The proposed method is implemented using Python with version 3.11.