ANALISIS CLUSTERING DATA PENYANDANG DISABILITAS MENGGUNAKAN METODE AGGLOMERATIVE HIERARCHICAL CLUSTERING DAN K-MEANS

The issue of disability is still an issue that attracts public perception in relation to the discrimination they encounter every day. Society and the environment still stigmatize or treat persons with disabilities unfairly. Even though many people with disabilities have abilities that are equivalent to people who do not have physical limitations. Through this case study, the researcher hopes to help contribute to raising awareness so that the rights of persons with disabilities can be fulfilled, particularly in relation to access to proper employment and education and to explore new information regarding persons with disabilities. by processing datasets of persons with disabilities in Indonesia by taking data samples in 7 provinces in Indonesia based on 137 cities and regencies with 3 types of disability variables namely physical disability, blind or deaf disabilities, deaf and speech disabilities. Using K-means algorithm clustering analysis and agglomerative hierarchical clustering. to identify outliers and anomalies researchers use the EDA (Exploratory Data Analysis) method in order to assist in enriching the results of data analysis. when determining the centroid the researcher used the agglomerative hierarchical method, namely the Average Linkage based on the validity test of the highest cophenetic correlation value, this produced the same results as in the previous study. a total of 62 cities and regencies, cluster 2 is a cluster with a moderate level category totaling 37 cities and regencies, cluster 3 is a cluster with a low level category totaling 27 cities and regencies. the best evaluation results used the DBI Davies bouldin-Index method by producing 2 clusters because the number of clusters with a lower Davies Bouldin-Index value was considered to indicate a better level of quality in cluster division.

Keywords: Persons with disabilities, agglomerative Hierarchical clustering, K-means , EDA (Exploratory Data Analysis), DBI Davies bouldin-Index