Claim Missing Document
Check
Articles

Found 1 Documents
Search

Comparison and Evaluation of Euclidean and Canberra Distances in the Adaptive K-Means Algorithm for Classifying the Food Security Status of Indonesian Provinces Cinthya Agatha Sinaga; Paska Marto Hasugian
Jurnal Komputer Indonesia (Ju-Komi) Vol. 3 No. 02 (2025): Jurnal Komputer Indonesia (JU-KOMI), April 2025
Publisher : SEAN Institute

Show Abstract | Download Original | Original Source | Check in Google Scholar | DOI: 10.58471/ju-komi.v3i02.752

Abstract

Food security issues in Indonesia are a major concern because they affect the sustainability of people's livelihoods and regional disparities. This study was conducted to classify food security conditions between provinces based on two main indicators, namely the Food Security Index and the Percentage of Adequate Food Consumption. The method used is the K-Means Adaptive algorithm with a comparison of two types of distance measurements, namely Euclidean and Canberra. The selection of centroids is done gradually using a probabilistic approach to improve the stability of the clustering results. Before conducting a comprehensive test, the method is first tested using sample data to see the characteristics of each distance function. Subsequently, all data were analyzed using Python programming, and the results were evaluated using the Silhouette Score metric. The analysis results showed that the Canberra distance function provided better clustering quality than the Euclidean function with a value of 0.415. This approach is expected to serve as a reference for more accurate and informative regional-based food security analysis.