Designing of Growth Reference Chart by Using Birespon Semiparametric Regression Approach Based on P-Spline Estimator

Abstract

The anthropometric measures i.e. weight and height are often used to measure physical children growth. So, for designing of growth reference chart, the simultaneous model estimation is better than partial single response model approach. The growth of children up to 12 months old grow rapidly, and than decrease slowly along with increasing of children age. It means that locally model approach more appropriate to the data. P-Spline Estimator is one of estimator in nonparametric regression. In this study we investigate the pattern of sex difference in growth reference chart of children from birth to five years old in Surabaya, Indonesia by using biresponse semiparametric model approach based on P-Spline estimator. The model consists of two response variables i.e. weight and height. Meanwhile, age and sex are as predictor variables. Based on children data in Surabaya city 2013, and the 50th percentiles estimation of weight and height, we obtain that in average the growth reference chart of boy higher than those of girl. In addition, the mean squared error value and coefficient of determination are 0.5684 and 99.99%, respectively. By considering both weight and height of reference growth charts, in Surabaya city, Indonesia, the nutritional status of girl is better than that of boy.