Key result
Exercise frequency decision tree predicts metabolic syndrome with ~86% accuracy in a population cohort.
Why the study?
Does exercise frequency and intensity correlate with the prevalence of metabolic syndrome components at the population level?
Observational (n=150)
Does exercise frequency and intensity correlate with the prevalence of metabolic syndrome components at the population level?
Answers are grounded in this paper and cite their sources.
Exercise frequency and intensity show complex ecological correlations with metabolic syndrome components, and a decision tree model can predict population-level MetS risk with high accuracy.
[Purpose] Metabolic syndrome (MetS) is a major risk factor for cardiovascular disease and diabetes. However, ecological evidence regarding the differential effects of exercise intensity remains limited. In this study, we examined the associations between three types of exercise (high-intensity, moderate-intensity, and strength training) and five MetS components at the population level and developed a decision tree-based model to identify high-risk groups.[Methods] Age group-level aggregate data (N = 150) from the National Health Screening Statistical Yearbook (2014-2023) were analyzed. Pearson’s and age-adjusted partial correlations were used to assess associations between exercise frequency and the prevalence of metabolic conditions. A CART-based decision tree model was developed and evaluated using five-fold stratified cross-validation.[Results] In simple correlation analyses, high-intensity exercise was negatively correlated with hypertension and hyperglycemia. After adjusting for age, moderate-intensity exercise was positively correlated with abdominal obesity, hypertension, and hyperglycemia, whereas strength training was positively correlated with hypertriglyceridemia and low HDL cholesterol (all p < 0.001). The decision tree model achieved an accuracy of 78.7% using high- and moderate-intensity exercise variables, which improved to 82.2% when all three exercise types were included. Under the MetS criterion (≥ 3 risk factors), accuracy reached 85.6% with a precision of 0.94.[Conclusion] Partial correlations differed in direction from findings at the individual level, indicating a potential ecological fallacy. Nevertheless, the decision tree model demonstrated acceptable predictive performance, suggesting that exercise frequency may be a useful population-level screening indicator.
No takes yet. Share an insight, caveat, or question.
Son et al. (2026) conducted an observational in Metabolic syndrome (n=150). Exercise frequency (high-intensity, moderate-intensity, and strength training) was evaluated on Prediction accuracy for metabolic syndrome (≥ 3 risk factors) using a decision tree model. A decision tree model using high-intensity, moderate-intensity, and strength training exercise frequencies achieved 85.6% accuracy in predicting metabolic syndrome at the population level.
Synapse has enriched 10 related papers. Consider them for comparative context: