Data Mining for Estimating the Impact of Physical Activity Levels on the Health-Related Well-Being

Abstract

The main objective of this paper is to employ some multifactorial data mining techniques for studying the direct and indirect effects of the physical activity intensity on persons’ health-related well-being. The availability of such a data-driven modeling and simulation interface enables analysts and decision makers to boost their decision by better understanding the types and levels of relationships between the main factors promoting the well-being of individuals. The data mining investigation is conducted at the CHU Gabriel Montpied (Clermont-Ferrand, France) on a population of employees, composed of medical and nonmedical staff. An observation-like study is conducted with the main aim of assessing direct and indirect effects of the physical activity intensity on the population’s health. This is especially performed by examining the significance of associations between physical activity indices and a set of their medical records. One of the main models resulting from data mining in this paper links cardiovascular risks to a set of exogenous variables including work and sport activity indices. The empirical results are consistent with many recent findings emphasizing the role of increasing high and intermediate levels of physical activity among health public-sector employees to effectively fight some diabetic and cardiovascular diseases.