Comparative Analysis of the K-Means Algorithm and the K-Medoid Algorithm in Clustering the Elderly Population

Abstract

The Central Bureau of Statistics is a Non-Ministerial Government Institution that is directly responsible to the President. So far, population data by sex and age group is done by calculating the population based on the age range starting from 0-4 years old up to 75+ which is separated by male and female sex (data attached). The data that has been processed so far is only a projection of population data, and is not reprocessed to obtain information about the grouping of population data by age. So Bengkulu City BPS has difficulty in mapping out the best program for the elderly population because Bengkulu City BPS does not yet have valid data on the number of elderly residents. The grouping of elderly population data in the K-Means Method and the K-Medoids Method was divided into 2 groups, namely high and low groups. Clustering based on male gender, the results of grouping in each clueter are different, while the number of iterations that occur is the same but the processing time between the two methods is different, where the K-Medoids method is faster than the K-Means method. Clustering based on female gender, the results of grouping in each clueter are different, while the number of iterations is the same but the processing time between the two methods is different, where the K-Medoids method is faster than the K-Means method. The results of the comparative analysis between the K-Means and K-Medoids method, it was found that the differences in the results of grouping, iteration and processing time occurred depending on the initial centroid value used in each method.