OPTIMASI METODE K-MEANS DENGAN ALGORITMA PSO PADA PENGKLASTERAN DATA BERDIMENSI TINGGI
Daftar Isi:
- K-means is one of the clustering methods whose clustering results depend on the position of the initial centroid. The initial centroid commonly used in the k- means method is the initial centroid randomly generated, initial centroids that are randomly generated usually often causes k-means to be trapped in the optimum local solution, therefore in this study we will examine the effect of particle swarm optimization which are wrong one optimization algorithm that can do a global search in determining the initial centroid of k-means. Clustering k-means with random initial centroids and initial centroids from particle swarm optimization calculations are each tested on data dimensional reduction and without dimensional reduction. Based on the results of evaluation of the results of k-means clustering with the initial centroid of particle swarm optimization able to improve cluster quality, both if tested on reduction and without reduction data, namely with the percentage change value of 43.8% in data without dimensional reduction and 53.4% in the data with dimensional reduction. Although it can increase overall computational time but, the initial centroid obtained from particle swarm optimization makes the complexity of k-means work simpler.