Leveraging Machine Learning Approach for Identification of Hypothyroid Disease

Abstract

Thyroid disease involves problems when diagnosing as it includes things such as hypothyroid and hyperthyroid conditions. Complications can be avoided if the problem is detected on time. This research presents XGBoost, Random Forest, Decision Tree, and SVM as machine learning (ML) techniques accompanied with feature selection techniques - RFE, UFS or PCA. Using datasets from UCI repository and clinical sources, up to 99.35% of the models reached classification accuracy. XGBoost was indicated as the best performer and optimized frameworks enhanced precision, recall and efficiency. The use of modern ML algorithms alongside feature selection techniques forms a cost effective and robust solution to the problem of accurate diagnosis of thyroid disease.

Ähnliche Arbeiten