Data Analysis and Related Applications, Volume 1

1Performance of Evaluation of Diagnosis of Various Thyroid Diseases Using Machine Learning Techniques

Thyroid cancer is the second most prevalent cancer type among women in Turkey. The number of people diagnosed with thyroid cancer in the United States in 2021 is estimated as 44,280, according to the report published by the American Cancer Society. The risk of thyroid cancer can be reduced by early diagnosis and treatment. This study is focused on predicting five different thyroid diseases, based on various symptoms and reports of the thyroid. Several machine learning algorithms, such as support vector machine, k-nearest neighbors, artificial neural network and decision tree are used for diagnosis of various thyroid diseases, and their classification performances are compared with each other. For this purpose, a thyroid disease dataset gathered from the Department of Nuclear Medicine and Endocrinology in Istanbul University-Cerrahpaşa Faculty of Medicine was used.

1.1. Introduction

According to Feigenbaum, the pioneer of Artificial Intelligence, “an expert system, which is one of the branches of Artificial Intelligence, is an intelligent computer program that uses knowledge and inference procedures to solve problems that are difficult enough to require significant human expertise for their solution” (Bursuk 1999). The first expert system is DENDRAL, developed by chemist Joshua Lederberg to describe chemical molecular structures in 1965. Since then, the spectrum of artificial ...