K-Means Bootstrap Analysis in the Birth Weight Classification of the Born Babies

Abstract

Low birth weight babies are caused by poor nutritional status before and during pregnancy. The impact of low birth weight is the slowing of the growth of infants seen in weight gain that does not reach normal levels when one-year-old. Classification of birth weight babies will be grouped into several groups, namely, low birth weight, normal baby weight, and excess baby weight. Classification of birth weight is done by cluster analysis. Cluster analysis is one of the statistical analyzes which aims to group objects based on the similarity of characteristics between those objects. Clusters of K-Means begin by determining in advance the number of clusters desired, so the measure of inaccuracy must be measured again. The inaccurate size of the K-Means cluster analysis can be done with the application of Clustering Bootstrap. This study aims to classify the birth weight of a baby using a cluster and a bootstrap cluster. The results showed a grouping of birthweight babies grouped into three clusters, cluster I had a baby's birth weight above 3130.171 kilograms, cluster II had a baby's weight under 3033.327 kilograms and cluster III had a baby's birth weight below 2299.994 kilograms. K-Means Bootstrap results produce relatively the same cluster. The best method is K-Means Bootstrap B = 75, with a total accuracy rate (TAR) of 0.7993 and a total error rate (TER) of 0.2007. K-Means Bootstrap can be used as an alternative way to determine the group of birth weight babies.