Cluster-Based Analysis For Growth Of Respiratory infectious diseases in children Using K-Means And Morran's I: Case Study Of North Aceh

Abstract

Respiratory infectious diseases in children are a major public health problem that requires in-depth understanding for effective treatment and prevention of the spread of these diseases. This research proposes a cluster-based analysis approach using the K-Means and Morran's I model to understand the growth of respiratory infectious diseases in children in the North Aceh district. This research aims to identify disease distribution patterns and group areas with similar risks to avoid spreading in adjacent areas.  The data used includes cases of respiratory infections in children for the last three years, starting from 2021 to 2023. The K-Means clustering model is used to group geographic areas that have similar patterns in the distribution of cases of respiratory infections in children. In contrast, Morran's I is used to measuring the level of spatial dependence between cases of respiratory infections in children in each region. The results showed that K-Means and Morran's I were able to identify spatial patterns of respiratory infectious diseases in children well. There are 3 clusters formed from the results of data categorization. Cluster 1 consists of 18 sub-districts characterized by having a small number of cases for all types of disease. Cluster 2 consists of 3 sub-districts characterized by a high number of cases of pulmonary tuberculosis, bronchitis, ARI, and a moderate number of cases of glandular tuberculosis. The cluster consists of 6 sub-districts characterized by high cases of pulmonary TB, low cases of glandular TB, and moderate cases of bronchitis and ispa. The spatial pattern formed from the Kmeans clustering results is in line with the results of Moran's I analysis, namely: cluster 1 has a dispersed pattern, and clusters 2 and 3 have a clustered pattern. Based on Moran's I analysis, cluster 1 is a cluster with negative spatial dependence on disease growth with an index value of -0.01091, while clusters 2 and 3 are clusters with positive spatial dependence on the growth and spread of respiratory infections with an index value of 0.675138 for each cluster 2 and 0.303281 for cluster 3.