Generalized Additive Models for Modeling Pneumonia Cases in Toddlers in West Java based on the Penalized Spline Estimator

Wahyu, Azkanul
Nurul Gusriani
Kankan Parmikanti

Eksakta : Berkala Ilmiah Bidang MIPA
Vol. 25 No. 02 (2024): Eksakta : Berkala Ilmiah Bidang MIPA (E-ISSN : 2549-7464)

Acute Respiratory Infections (ARI) are one of the causes of high mortality in the world, such as pneumonia in toddlers. Pneumonia cases in West Java are high compared to other provinces. In this study, pneumonia cases will be modeled with Generalized Additive Models (GAM) based on penalized spline estimators. The optimal number of knots is determined using the full search algorithm and the optimal smoothing parameter is obtained based on the minimum Generalized Cross Validation (GCV) value of order one or two. Then, GAM parameter estimation is performed using the local scoring algorithm. Formed model based on the order, number of knots, and smoothing parameters of each predictor variable with order one, number of knots two, and optimal smoothing parameter one for , order two, number of knots three, and optimal smoothing parameter one for , and order one, number of knots two, and optimal smoothing parameter for  whose parameters were estimated by local scoring resulted in a coefficient of determination of 0.679. This indicates that 67.9% of the factors from the predictor variables affect the percentage of pneumonia cases among under-fives while the remaining 32.1% is influenced by other factors outside the model.