Optimized Classification and Prediction with Hybrid Machine Learning Models

In recent years, there has been an increase in thyroid disorders. Early thyroid illness identification is a crucial endeavor because of its significance in metabolism. Although there are numerous studies on the identification of whether thyroid illness leads to cancer or not, in order to properly address and handle this condition, a timely and accurate medical diagnosis is crucial. Thyroid cancer detection and diagnosis could be enhanced by algorithms that utilize machine learning, which have achieved considerable amounts of interest especially in the medical field. This paper describes how machine learning techniques are being used to diagnose thyroid cancer. K-Nearest Neighbor, Decision Tree, Random orest, Ada-Boost with decision tree classifier, Gradient Boosting Classifier, Stochastic Gradient Boosting classifier, Extended Gradient Boosting Classifier, Extended Gradient Boosting Classifier with hyper parameter tuning, Extra Tree Classifier, Light GBM Classifier, Voting Classifier are among the machine learning approaches that are used and evaluated to assess how well they diagnose thyroid Cancer. The study evaluates the methods precision, recall, F1-score, and Accuracy. From all the above machine learning models Random Forest, Ada Boost, Extended gradient boosting classifier, Extra Tree Classifier, Light Gradient Boosting Machine outperforms the other models.