Thyroid Disease Prediction

Authors:

D. Yuvaraju1, V. Raghu, B.Shashank3, G. Jagadeesh4, Prof. Dr Rahuul Paatil 4

1,2,3 CSE Department of Computer Science Engineering(AI&ML), Sandip University, Nashik, Maharashtra, India.

⁴ CSE Department of Computer Science Engineering(AI&ML), Sandip University, Nashik, Maharashtra, India.

***

Abstract:

Thyroid disease prediction is a critical task in modern healthcare systems, enabling early diagnosis and effective treatment planning. Accurate identification of thyroid disorders such as hypothyroidism and hyperthyroidism can significantly reduce long-term health complications. This paper presents a machine learning-based thyroid disease prediction framework combining clinical, biochemical, and symptomatic data. The proposed system utilizes structured patient data including Thyroid Stimulating Hormone (TSH), Triiodothyronine (T3), Thyroxine (TT4), Free Thyroxine Index (FTI), age, sex, and symptom indicators. Multiple machine learning algorithms including Logistic Regression, Random Forest, and Extreme Gradient Boosting (XGBoost) are evaluated. The XGBoost model achieves superior performance with an accuracy of 96%, outperforming traditional classification models. The system is deployed as a web-based application using Flask, supporting both single patient input and batch CSV predictions. Experimental evaluation and performance analysis confirm the robustness, scalability, and reliability of the proposed approach in real-world medical applications.

KeyWords: Thyroid disease prediction, machine learning, XGBoost, healthcare analytics, clinical data, classification.