Open AccessImprove Hybrid Particle Swarm Optimization and K-Means for Clustering
Author(s) -
Yudha Alif Auliya,
Wayan Firdaus Mahmudy,
Sudarto Sudarto
Publication year - 2019
Publication title -
jitecs (journal of information technology and computer science)
Language(s) - English
Resource type - Journals
eISSN - 2540-9824
pISSN - 2540-9433
DOI - 10.25126/jitecs.20194183
Subject(s) - silhouette , cluster analysis , particle swarm optimization , centroid , computer science , k means clustering , swarm behaviour , cluster (spacecraft) , mathematics , data mining , mathematical optimization , artificial intelligence , programming language
. Potato production is strongly influenced by the selection of suitable land for crops. Criteria for land suitability of planting potatoes is influenced by climatic factors and land characteristics. planted area clustered based on 11 criteria land suitability. The clustering results in the form of four clusters, namely: very suitable (S1), appropriate (S2), is quite suitable (S3) and are not suitable (N). Clustering of land aims to improve the quality and quantity of the potato crop. Clustering is done using a hybrid Particle Swarm Optimization with K-Means (KCPSO). The hybrid method is used to obtain an accurate result cluster. In this study used a new approach to doing improve KCPSO with random injection method. The calculation of the value of cost based on the silhouette coefficient. The results obtained KCPSO showed better results when compared to using the K-Means algorithm without hybrid. The calculation result KCPSO get the best centroid indicated by the value of the largest Silhouette coefficient.Keywords: Clustering, K-Means, Particle Swarm Optimization, random injection, silhouette Coefficient.