A ORDINAL LOGISTIC REGRESSION BAGGING FOR MODELING AND CLASSIFICATION OF THE NUTRITIONAL STATUS OF TODDLERS IN SOUTHEAST PONTIANAK SUB-DISTRICT

Abstract

Although Pontianak's 2022 stunting rate of 19.7% is higher than the RPJMN's 2020–2024 target of 14%, this is still significant. The categories of stunts are very short (severely stunted), short (stunted), normal, and high, based on a high index of the body by age (TB/U). Ordinal Logistic Regression is one classification that can be used to group stunts based on the TB/U index. This approach makes the unstable parameter. Use the bagging to get stable parameters. The study aims to model and classify toddlers' nutritional status using the TB/U index. Utilizing secondary data for 150 toddlers from Pontianak Tenggara's UPT Puskesmas Parit Haji Husin II, This will monitor kids' growth from 24 to 59 months in 2022. Response factors include short, very short, normal, and high. The mother's job position, birth weight, length, and gender are the predictive variables. Due to imbalanced data utilized in the first analysis using Ordinal Logistics Regression, a decent model, and the final classification result, they used the Bagging OLR ensemble method. The study's findings are a very effective model using OLR Bagging, with an accuracy rate of 99.33%, a sensitivity value of 98.91%, and a specificity value of 98.52%. The results also revealed significant variables that influence the mother's employment status and the birth length variable.