Pemodelan Regresi Nonparametrik Dengan Pendekatan Spline Pada Kasus Stunting Di Indonesia

Abstract

Stunting is a condition of failure to grow due to chronic malnutrition which is caused by inadequate nutritional intake for a long time, resulting in being shorter than standard. Indonesia is ranked fourth in the world and second in Southeast Asia with the stunting rate in Indonesia reaching 24.4% or around 23 million children in 2021, while in 2022 the figure will drop to 21.6%. This research aims to determine several factors that influence stunting in toddlers in Indonesia using the nonparametric spline regression method with one node, two nodes and three nodes. This method is used because it can explain the pattern of relationships between predictor variables and response variables that do not follow a certain pattern and changes in behavior at certain intervals. The best model obtained using nonparametric spline regression is a one-knot model. The results of nonparametric spline regression modeling with the best one knot model are GCV of 14.32605 and 𝑅2 of 81.1%. Of the five independent variables, namely: toddlers receive complete basic immunization(𝑥1), babies receive exclusive breast milk for 6 months (𝑥2), babies born receive IMD (𝑥3), children aged 6-23 months consume five of these eight variables. food and drink groups throughout the day (𝑥4), households have access to proper sanitation (𝑥5), there are variables that do not have a significant effect, namely complete basic immunization (𝑥1).