An Intelligent Machine Learning Model for Early Detection of Thyroid Disease

Abstract

Thyroid disease is a prevalent health condition affecting a significant portion of the global population, and early detection is critical for effective treatment. This study focuses on improving thyroid disease prediction by utilizing machine learning (ML) models, specifically the XGBoost algorithm, in combination with hyperparameter optimization techniques. The goal of this work is to develop a robust predictive model that can accurately classify patients based on their demographic and medical data, such as age, gender, and thyroid hormone levels. The optimization of hyperparameters allows the model to fine-tune its performance, achieving higher accuracy and reducing the training time and memory requirements. The results indicate that the proposed model outperforms traditional ML models like Decision Tree, KNN, Random Forest, and Support Vector Machine, making it a promising solution for early detection and classification of thyroid disease. In the proposed framework, the XGBoost algorithm, known for its powerful prediction capabilities and wide range of customizable hyperparameters, is employed to categorize patients across multiple severity levels of thyroid-related diseases. The optimization process involves identifying the best hyperparameter values to maximize model performance. The model is trained and evaluated on a shared dataset, and its accuracy is compared with other ML models to ensure its superiority. The results demonstrate that the proposed XGBoost-based model achieves the highest accuracy, precision, recall, and F1 score, proving its potential as an effective tool for thyroid disease prediction. By optimizing hyperparameters and fine-tuning the model, the study presents a promising step towards more accurate, reliable, and efficient healthcare diagnostics.