Knowledge discovery using clustering methods in medicine: A case study for reflux disease

Abstract

Digitalization spreads day by day around the world; thus, the amount of data collected is on the rise. An increasing amount of data leads us to use the data and get the advantage of it by using methods like Data mining. Data mining is used in several industries. Especially as medical data is essential to be understood, it is crucial to work on it. Reflux disease is a painful illness spreading around the world. Reflux is more common compared to formerly known numbers of patients. Even though reflux is not as fatal as cancer, it decreases the quality of life and makes many people suffer in their daily life. So, reflux is affecting mental health directly. If we can ease the process of diagnosis of reflux, we may provide a better quality of life for people. In this study, various data mining algorithms are applied, and it is seen from results that medical care can be improved by changing. Nowadays, artificial intelligence applications in the field of gastroenterology stand out in various sources in the literature. However, a large database required that is specific for Reflux disease to implement these applications is available only at the Reflux Research Center in Ege University in Turkey. By benefiting the Short Form36 and Quadrad12 questionnaire data in this database, 3,909 patients and many artificial intelligence algorithms were used to discover the hidden associations among responses in the quality of life of these patients. The algorithms used in the tests are Apriori, Frequent Pattern Growth, Density-Based Spatial Clustering of Applications with Noise, Self-Organizing Map, and KMeans. In the tests, it was observed that the most successful algorithm in terms of the structure of the data was KMeans, and a set of remarkable 27 rules according to the optimal Sum of Square Error value was obtained. © 2025 Elsevier B.V., All rights reserved.