Bagian 1: Kombinasi Metode Klastering dan Klasifikasi (Kasus Pandemi Covid-19 di Indonesia)

Abstract

The purpose of this research is to combine the classification and classification methods that are part of data mining. The case raised was the number of the spread of the Covid-19 pandemic in Indonesia as of July 7, 2020 with 34 records. Data sources were obtained from Ministry of Health Data, sampled and processed from covid19.go.id and bnpb.go.id. The variables used in the study are the number of positive cases (x1), number of cases cured (x2) and number of deaths (x3) by province. The classification and classification methods used are k-medoids and C4.5. The k-medoids method works to map clusters of regions in Indonesia by province. The mapping labels used are 3 clusters: high cluster (C1 = red zone), alert cluster (C2 = yellow zone), low cluster (C3 = green zone). The results of the mapping are continued using the C4.5 method to see the rules in the form of a decision tree. The analysis process is assisted with the RapidMiner software. Determination of the number of clusters (k) is determined by using the Davies Bouldin Index (DBI) parameter to optimize the cluster results obtained. For k = 3 has an optimal value of 0.740. The mapping results obtained 9 provinces are in the high cluster (C1 = red zone), 3 provinces are in the alert cluster (C2 = yellow zone) and 22 provinces are in the low cluster (C3 = green zone). The value obtained from the decision tree for cluster height (C1 = red zone) based on C4.5 is if the number of positive cases is smaller than 9524 and greater than 4329 (4329> x1