Optimization of the K-Means Clustering Algorithm Using Davies Bouldin Index in Iris Data Classification

Devanta Abraham Tarigan Devanta (STMIK Kristen Neumann Indonesia, Medan)

KLIK: Kajian Ilmiah Informatika dan Komputer
Vol. 4 No. 1 (2023): Agustus 2023

Data grouping is done by calculating the shortest distance to the initial cluster center point as the central point in the formation of each group or cluster. The results of the K-Means optimization study with the Davies Bouldin Index k-means clustering by dividing the cluster values 3,5,7,9. In testing the K-3 cluster values, the performance value of the average centroid distance is -0.312, then the K-Means optimization results with Davies Bouldin have a performance percentage of -0.799. Testing the value of the K-5 cluster has a performance value of an average centroid distance of -0.310, then the K-Means optimization results with Davies Bouldin have a performance percentage of -0.806. Testing the value of the K-7 cluster has a performance value of an average centroid distance of -0.310, then the K-Means optimization results with Davies Bouldin have a performance percentage of -0.806. Testing the K-9 cluster values has a performance value of an average centroid distance of -0.310, then the results of the K-Means optimization with Davies Bouldin have a performance percentage of -0.806. From the test results with variations in cluster values 3,5,7,9 it can be concluded that the optimization of the K-Means method with the Davies Bouldin Index testing the K-3 cluster values has better performance with an average value of -0.312 centroid distance then the results of K-optimization Means with Davies Bouldin has a performance percentage: -0.799.