Spatial-Temporal Visualization of Tuberculosis Vulnerability in East Java, Indonesia, Using the K-medoids Clustering Algorithm

Abstract

East Java is one of the provinces in Indonesia with a high tuberculosis rate. East Java has the second-highest rate of tuberculosis cases in Indonesia, and it still has a significant health issue that has to be resolved. An overview of the correlation between geographic locations and susceptibility levels can be obtained from tuberculosis susceptibility mapping. Using spatial analysis, community-based tuberculosis initiatives can be planned more successfully by identifying the pattern of illness distribution and potential causes. The K-medoids clustering technique is used in this paper to provide a new approach for mapping the degree of tuberculosis susceptibility in East Java. The overall number of TB cases, unhealthy houses, population density, and health facilities are all utilized as variables that affect vulnerability. These parameters are highly correlated with one another to assess the level of TB susceptibility. There are three categories for tuberculosis susceptibility levels: low, medium, and high from 2016 to 2020. According to the K-medoids clustering 2016-2020 data’s average silhouette, there are 0.445 clusters with a number of 3, 0.423 clusters with a number of 4, and 0.342 clusters with a number of 5. The silhouette value demonstrates that using 3 clusters improves the accuracy of the K-medoids computation. The average variance for the K-medoids approach is 0.041, whereas the average variance for the K-means method is 0.048. This demonstrates that the K-medoids algorithm is greater to K-means at forming clusters.