Sick Euthyroid Detection Using Machine Learning Techniques

People who are suffering from a chronic disease and have an imbalance in their thyroid hormone levels are experiencing a condition known as Euthyroid Sick Syndrome (ESS). ESS is frequently seen in patients with severe systemic illnesses, where thyroid hormone levels are disrupted, though the thyroid gland itself remains functionally intact. In the past many years, machine learning (ML) has played a role as an effective mechanism in the area of medical research to improve the understanding of diagnosis and treatment of complex disorders such as ESS. Providing early and precise diagnosis serves to avoid misdiagnosis and makes sure that patients are not facing unsuitable medical procedures. By developing predictive models and analyzing complex clinical data, machine learning can improve the accuracy of ESS, facilitate early detection, and ultimately enhance patient outcomes. This paper aims to develop an ensemble ML-based model with 5-fold cross-validation to classify Sick Euthyroid disease. Random Forest (RF) and Support Vector Machine (SVM) are used as base models, and Logistic Regression is used as a meta-model. We individually apply several other classifiers like Decision Tree, SVM, Logistic Regression, and RF and compare the performance. The stacking classifier gives a better accuracy of 99.35% among these techniques.