Machine Learning based Predictors of Cardiovascular Disease among Young Adults
Abstract
The purpose of this research was to develop a data-driven model to test the association of physical, metabolic, and hemodynamic variables on the risk of cardiovascular disease. The Structural Equation Modelling using Partial Least Square Method has been adopted to analyze the data. A sample size of 685 young adults who were in sedentary, physically trained, and endurance tested categories has been used in this research. Results have revealed that age and weight were the prominent predictors of the cardiovascular disease among the physical variables, total glucose and triglycerides were the prominent predictors among the metabolic variables, and systemic vascular resistance and systolic blood pressure were the prominent predictors of the cardiovascular disease among the hemodynamic variables. It was concluded that while all the three variables which are considered to be the antecedents of risk of cardiovascular disease, not all the parameters listed under these three categories have a statistically significant influence on the risk of the cardiovascular disease. The results can be of use to the medical practitioners as well as researchers in machine learning, as it adds to the repository of earlier studies and can be used by the medical professionals in effective decision making in disease prediction.
