Identifying and clustering of COVID-19 cases in Jakarta using K-medoids algorithm

Abstract

Indonesia has been hit by an outbreak of the corona virus (COVID-19). Every day the spread of COVID-19 in Indonesia continues to increase. People practice social distancing to break the chain of the spread of COVID-19 which is spreading in various regions in Indonesia. Many people have experienced losses and difficulties in their MSMEs due to COVID-19. However, when the COVID-19 pandemic emerged, product marketing management experienced a decline because there was no smooth operation in the distribution process for product delivery which was not carried out normally. Therefore, it is necessary to formulate a marketing management strategy that can reveal the condition of product sales for MSME craftsmen. People can contract COVID-19 from other people who are infected with this virus. Therefore, it is necessary to explore the infected area, so that the hotspot areas can be guarded by both the government and citizens. Therefore, research is needed on the spread of the COVID-19 virus in DKI Jakarta. The ultimate goal of this research is to create a cluster system to classify areas affected by COVID-19 by applying the yahoo K-medoids method, which is well-known for its accuracy. To perform clustering, data for the DKI Jakarta City area infected with COVID-19 were taken from a web portal provided by the DKI Jakarta Provincial Government. In the implementation of the K-medoids algorithm, a clustering simulation (grouping) of the DKI Jakarta area that is infected with COVID-19 is carried out using the rapid miner application. From the results of clusters that have been tried, the optimal BDI value was obtained in clusters in simulation 4 (five clusters) and a value was obtained of 0.263.