JoWUA

Abstract

Machine learning uses different symptoms of patients for the disease prediction purpose. There are many repositories which provide the symptoms information in the form of datasets. Machine learning is capable of predicting the disease automatically by analysing the dataset. Accuracy of prediction depends on various factors such as features available in datasets and ML model used. The usage of ensemble models for classification problems is examined in the paper. UCI repository is used to extract the thyroid dataset and subjected to a number of machine learning methods, including Naive Bayes, Support Vector Machines (SVM), K-Nearest Neighbours (KNN), Logistic Regression, and Stochastic Gradient Descent (SGD). Metrics including as accuracy, precision, recall, and F1-score are used to assess model performance. Additionally, the impact on model performance is seen when choosing filter features using various approaches such as Information Gain, Gini Index, Chi Square, and Relief F. Finally, to enhance the performance of the model, an ensemble model is created using the AdaBoost approach and the GridSearchCV optimization algorithm.