Identification of risk group and factors affecting metabolic syndrome in one-person households: Machine learning and latent class analysis approach (Preprint)

Abstract

The rapid increase in one-person Korean households has led to an outbreak of metabolic syndrome. This calls for an analysis of the complex effects of metabolic syndrome risk factors in one-The present study aimed to identify the factors affecting metabolic syndrome in one-person households using machine-learning techniques and categorically characterize its risk factors through latent class analysis.person households, which vary from individual to individual.

The present study aimed to identify the factors affecting metabolic syndrome in one-person households using machine-learning techniques and categorically characterize its risk factors through latent class analysis.

This cross-sectional study included 10-year secondary data of the National Health and Nutrition Survey (2009–2018). We selected 1,371 participants belonging to one-person households. Data were analyzed using SPSS 25.0 (IBM, New York), Mplus 8.0 (Muthen & Muthen, Los Angeles), and Python 3.0 (Plone & Python, Montreal).

Machine-learning techniques investigated the factors affecting metabolic syndrome in one-person households. We categorized the metabolic syndrome risk factors in one-person households hierarchically into four classes. Results showed that those with obesity and abdominal obesity in middle adulthood exhibited the highest probability, indicating that they are the most vulnerable and at-risk group (P < .001).

This study identified the factors affecting metabolic syndrome in one-person households using machine-learning techniques and latent class analysis. Customized interventions prepared for each risk factor for one-person households can prevent metabolic syndrome.