Peningkatan Ketepatan Klasifikasi dengan Metode Bootstrap Aggregating pada Regresi Logistik Ordinal

Abstract


 Baby's birth weight is influenced by characteristics of pregnant women such as age, parity, education level, pregnancy visit, and gestational age. Classification of the birth weight of a baby is grouped into several groups, namely low birth weight babies, normal baby weight and excess baby weight. The classification method with ordinal logistic regression provides an unstable parameter estimation, which means that if there is a change in the data set causes a significant change in the model. So that to obtain a stable parameter estimation in the ordinal logistic regression model is used aggregating (bagging) bootstrap approach. This study aims to improve the classification of ordinal logistic regression by using bagging on a baby's birth weight. The classification results with bagging ordinal logistic regression were able to reduce classification errors by 20.237% with 76.67% classification accuracy