Klasterisasi Menggunakan Algoritma K-Means dan Elbow pada Opini Masyarakat Tentang Kebijakan Sekolah Luring Tahun 2022

- The covid-19 pandemic that swept across the globe had adverse effects in many areas. One of the most affected areas is education in Indonesia. The online learning model became the only option at the time, which had a negative impact on the quality of education in Indonesia. As time went on, conditions are getting better, but there was still a threat of covid-19. In early 2022 governments began to adopt face-to-face or offline learning that attracted opinions on social media. The opinions that are widely written on social media need to be prepared because they could be input to the government. Clustering using the k-means algorithm with the elbow method as its optimizer in determining the best cluster number is one of the opinions processing options on social media for measuring and accounting. Data is treated with two approaches: with and without stemming . Applying the elbow method to the k-means algorithm produces a performance of the clustering model with a DBI value of 0.003 with 4 clusters, and a value of SSE 0.331, for data without stemming . On data with treatment using stemming , it has 3 cluster numbers with a value of DBI at 0.003 and SSE at 0426.