Course clustering in Moodle based learning management system using unsupervised learning

The covid-19 pandemic has not shown a significant decline in the graph, entering the beginning of 2021 the spread of covid-19 tends to continue to increase in several areas. This forces universities in Indonesia to conduct lectures online using a platform that suits the needs of the higher education institution. This refers to the Joint Decree of the Minister of Education and Culture (Mendikbud), Minister of Religion (Menag), Minister of Health (Menkes), and Minister of Home Affairs (Mendagri) regarding Guidelines for Implementation of Learning in the Academic Year 2020/2021 and Academic Year 2020/2021. during the Pandemic Coronavirus Disease 2019 (Covid-19). Moodle is a platform that is quite widely used by universities in conducting lectures. Moodle has complete features that can replace lecturing activities that have been carried out offline to online. In addition, Moodle also has a datalog that records all online learning activities for lecturers and students so that learning data can be analyzed and audited. One way to analyze online learning using Moodle is by clustering courses, especially on learning activities that have been carried out by lecturers and students that are recorded in the datalog. By clustering courses, it is hoped that the character of online learning courses can be seen as the initial basis for taking further action to maintain the quality of online learning at the institution. The result of this study indicates that the performance of online lectures in Virtual Class University of Lampung using K-Means according to the amount of hit per user and amount of hit per resource divided into 5 clusters, namely cluster 1 with 32 courses in Active status, cluster 2 with 69 courses in Normal status, cluster 3 with 6 courses with Onfire status, cluster 4 with 10 courses in More Active status, and cluster 5 with 225 courses in Pasif status.