Association Analysis of COVID-19 Outbreak in Thailand Using Data Mining Techniques

Abstract

The objective of this research is to apply data mining techniques to determine the relationships that affect the COVID-19 outbreak among infected people in Thailand. The data mining techniques included cluster analysis using K-means clustering and association rule mining. Cluster analysis was used to classify the infection cases into the appropriate group. Association rule mining was applied to find the relationship between datasets to find patterns associated with each group. The infection causes were obtained from the online data source of the Department of Disease Control, Ministry of Public Health, Thailand. The datasets consisted of 103,639 cases that occurred during the second wave and the third wave (December 2020-May 2021) of the outbreak. The outbreak had the largest number of infected cases and was the most severe since the epidemic occurred in Thailand during the past year. The variables selected for data analysis were age, sex, province of onset, and infection sources. The results from clustering consisted of 7 groups, which were categorized by age group ranging from 0-9 years to 60 years or over. Next, the association rule mining was applied to find the co-occurrences between itemset. The 128 rules were generated, then illustrated the results based on the value of support, and lift by using network visualization. The association network provides the major causes and effects in each group, which can be used as a precaution to prevent the outbreak in the future.
Keywords: