0.252
P-Index
This Author published in this journals
All Journal Biometrika dan Kependudukan
Bambang Wijanarko
Fakultas Kesehatan Masyarakat, Universitas Airlangga
Articles
1
Documents
Penerapan Clustering Bootstrap dengan Metode K-Means

Biometrika dan Kependudukan Vol 3, No 1 (2014): Jurnal Biometrika dan Kependudukan
Publisher : Biometrika dan Kependudukan

Show Abstract | Original Source | Check in Google Scholar | Full PDF (115.545 KB)

Abstract

ABSTRACT Cluster analysis was a process for grouping a set of objects based on data that have similarcertain characteristic. K-Means was a method of cluster analysis which begins by determining the number of clusters desired. Bootstrap was a sampling technique with replacement from the original sample. Bootstrap was used to estimate the parameters based on minimal data using a computer. This methode was useful to maximize relative diffrence and variation in the clusters. Malnutrition was a major problem in Indonesia and is still a concern in children under five. Infants with malnutrition would have a higher mortality rate. The purpose of this study wasto assess the accuracy of K-Means and Bootstrap K-Means method to clustering nutritional status of children undersfive which was crosstabulated with the nutritional status of children based on the WHO-2005 in the Ajung Public Health Center, Jember. The variable in this study was nutritional status based on WHO criteria 2005 as standard benchmarks, presentage and weight. This was non-reactive research, using secondary data in Ajung Public Health Center, without any direct interaction with the subject. This study concluded that the total accuracy rate (TAR) and Total Error Rate (TER) to determine nutritional status of  K-Means method was TAR=0.9 and,  TER=0.1; Bootstrap K-Means methode (B=25) TAR=0,925 and TER=0.075; Bootsstrap K-Means methode (B=50) TAR=0.9417, TER=0.0583;and Bootstrap K-Means Bootstrap (B=75) TAR=0.9583 and TER=0.0417 after crosstabulated with nutritional status based on WHO-2005 (weight for age). In conclusion general, the K-Means method and Bootstrap K-Means method and crosstabulated with nutritional status based on WHO-2005 has shown very good accuracy to determine the nutritional status of children. The best method was Bootstrap K-Means (B=75). K-Means Bootstrap methods can be used as an alternative way to determine the nutritional status of children. Keywords: cluster analysis with K-Means method, bootstrap, nutritional status