Data Analytics for Metabolic Syndrome Diagnostics

Abstract

Metabolic syndrome (MS) represents an important risk factor for the development of cardiovascular diseases, as well as type 2 diabetes mellitus, which as one of a few clinical syndromes affects more than 25% of the world population. The diagnosis is often associated with various negative activities like little physical exercise, poor diet, stress, genetic predisposition, and excessive alcohol consumption. The aim of this paper is to provide a literature review of the current state of the art in the area of MS diagnosis by means of data mining methods. We structure our literature review by means of the CRISP-DM methodology, which is typically used to organize the analytical process. The reviewed problem was most often approached as a binary classification problem and frequently used methods have been decision trees, neural networks and logistic regression. Some of the authors applied also suitable statistical methods like Welch’s t-test, Pearson’s chi-squared test. Mostly, the size of analyzed data samples was more than one thousand patients.