The article compares different regression models to predict body fat percentage in university male students using factors like height and weight. The study found that the multivariate regression model and the model based on body surface area gave similar results, with one model being superior.