K-means and fuzzy c-means algorithm comparison on regency/city grouping in Central Java Province

Abstract

The Human Development Index (HDI) is very important in measuring the country's success as an effort to build the quality of life of people in a region, including Indonesia. The government needs to make groupings based on the needs of a city/district. To facilitate data grouping based on the similarity of existing characteristics, it is necessary to have a data grouping method, namely the clustering technique. There are several algorithms that are often used in clustering techniques, namely K-Means and Fuzzy C-Means. Each algorithm has advantages and disadvantages. Therefore, in this research, the comparison of the best clustering algorithms will be discussed. The purpose of this research is to compare the K-Means and Fuzzy C-Means algorithms in the grouping of Regencies/Cities in Central Java Province in 2021 based on the Human Development Index (IPM) indicator. The method used is descriptive qualitative. The data used is obtained from the Central Java Statistics Agency, which is in the form of the Central Java Province HDI indicator in 2021 which consists of 35 Regency/City members and has 4 variables, namely Life Expectancy (AHH) (X1), Expectation of School Years (HLS) (X2), Average Length of School (RLS) (X3) and Expenditure Per Capita (X4). The results showed that after a comparative analysis with the standard deviation ratio method, the FCM clustering method was better than the K-Means method. The value of the FCM standard deviation ratio is 0.460093 and the K-Means standard deviation ratio is 0.473601.

Keywords

• k-means
• fuzzy c-means
• clustering