Robust clustering using sparse K means on text analitycs of PPKM policy in Indonesia

Imposes Restrictions Towards Community Activities (PPKM) Policy is the latest policy issued by the government to deal with the problems of the pandemic in Indonesia. The design and implementation of the PPKM Policy is carried out with the main objective being for the benefit of health sector and economy. PPKM policy is currently one of the topics that are often discussed by community on Twitter social media. The response of twitter user's in the form of unstructured data to the implementation of PPKM policy can be grouped using Robust Sparse K Means clustering (RSKM). This algorithm is a combination of the sparse K-means and Trimmed K-Means algorithms, and it can handle large datasets and contain outliers especially on Text Analytics. Data used is obtained through the user's twitter account via twitter API based on the keywords #PPKM LEVEL 4 and #PPKM JAWA BALI. The study shows that Robust Sparse K-Means analyzing found that 10 clusters with the best parameter of Classification Error Rate value compared to other algorithm, as well as the content most discussed were the extension of PPKM carried out by the government in Java and Bali even though the level had dropped