Penerapan Metode KnA (Kombinasi K-Means dan Agglomerative Hierarchical Clustering) dengan Pendekatan Single Linkage untuk Menentukan Status Gizi pada Balita

Authors

  • Intan Alpiana Jurusan Teknik Informatika, Teknik Informatika, Universitas Negeri Surabaya
  • Lilik Anifah Jurusan Teknik Informatika, Teknik Informatika, Universitas Negeri Surabaya

DOI:

https://doi.org/10.26740/inajet.v1n2.p61-68

Keywords:

Data Mining, Agglomerative Hierarchical Clustering, K-Means, Nutritional Status Of Toddlers.

Abstract

The nutrition fulfillment for children under five years old is a factor that must be considered in maintaining their
health, because toddlers are in a developmental period that is vulnerable to nutrition. The attention of parents and officers of
the Public Health Service Center (PUSKESMAS) is very necessary to monitor toddler nutrition. Therefore we need a system
that can help to classify the nutritional status of children. The system is designed using the KnA method (a combination of
K-means and Agglomerative Hierarchical Clustering) so that clustering results are better by combining the advantages of
bottom-up clustering (agglomerative) and top-down clustering (k-means). This study attempts to categorize the data from
Pulung Ponorogo health center with 7336 data (testing 3641 and learning 3695) with variables of sex, age, weight and
height into 4 clusters of each toddler's nutritional status, namely for BB / U there 4 clusters BB S.Krg, BB Krg, BB Normal,
BB Lbh, 4 clusters TB / U consisting of S. Short, Short, Normal, High and BB / TB 4 clusters S.Kurus, Skinny, Normal and
Fat. In this study, the KnA method was able to properly classify toddlers' data in large numbers. The learning data is as
much as 3695 produces an accuracy of BBU = 94%, TBU = 98%, and BBTB = 93%. From the results of the grouping
conducted, it is used to determine the nutritional status of status according to BB / U, TB / U, and BB / TB, each of which i s
equipped with a handling solution. From the testing with the black-box method as a whole the system is running well, as
well as the questionnaire method and as well as validation of nutri

Downloads

Published

2022-03-17

Issue

Section

Articles
Abstract views: 413 , PDF Downloads: 1487