Classification of Thyroid Class using ID3 Algorithm and Artificial Neural Network (ANN)

Abstract

Thyroid disease refers to a range of conditions or issues affecting the thyroid gland. This gland, located below the Adam’s apple, is responsible for coordinating various metabolic processes in the body, making its function essential. Early detection of thyroid symptoms is crucial as an initial step in planning the necessary treatments to prevent more severe thyroid-related health risks. One commonly applied method for early detection involves classification using a data mining approach. Among the algorithms frequently used for classification are the ID3 algorithm and Artificial Neural Networks (ANN). This study aims to obtain the best classification results for detecting thyroid disease by comparing these two algorithms. The accuracy results for percentage split testing were 88% for ID3 and 90% for ANN. Meanwhile, the accuracy values for K-Fold cross-validation were 93% for the ID3 algorithm and 95% for the ANN algorithm. Additionally, the overall average precision and recall values for both algorithms were above 75% for percentage split testing and above 90% for K-Fold cross-validation. The results indicate that ANN achieved higher percentages compared to ID3. Based on the accuracy, precision, and recall values obtained from both algorithms, it can be concluded that the ANN algorithm performs better than ID3 in classifying thyroid disease.

Keywords

• thyroid disease, classification, data mining, id3 algorithm, ann algorithm