The rapid development of technology impacted the world of education worldwide, including in Indonesia. Good problem-solving skills are needed in this era. Computational thinking (CT) or computational thinking is considered capable of training students in problem-solving skills. This study aims to group students based on CT abilities to help teachers more easily determine learning methods that suit the characteristics of students. The data processing uses the K-Means algorithm with data taken from the results of the Bebras Challenge 2022 for the elementary school level. In the clustering process, the most optimal number of clusters was determined using the Elbow method. The optimal cluster is 3 clusters: namely high, medium and low levels of understanding of CT. The cluster for the SiKecil category has an average value of 81.52 and a duration of 15.07 minutes for the high cluster, an average value of 43.02 and a duration of 24.59 minutes for the medium cluster, and an average value of 34.96 and a duration of 15.28 minutes for the low cluster. The clusters formed for the Siaga category are clusters with a high level of understanding of CT with an average value of 67.13 and a duration of 23.88 minutes, an average value of 56.91 and a duration of 34.76 minutes for the medium cluster, and low cluster with an average value of 32.14 and a duration of 20.79 minutes.
Copyrights © 2023