Pemetaan Kesulitan Belajar Mahasiswa Menggunakan Algoritma K-Means Clustering

This study aims to analyze the level of learning difficulties among students of the Information Systems Study Program at Nusa Putra University based on their perceptions of the learning process. Data were collected using a Likert-scale questionnaire covering variables of material understanding, learning methods, learning interaction, learning media, and learning environment. The analysis was conducted using an Educational Data Mining approach by applying the K-Means Clustering algorithm, while the optimal number of clusters was determined using the Elbow Method. The results indicate the formation of three learning difficulty clusters, namely low, moderate, and high, with most students classified in the moderate learning difficulty cluster. This condition suggests that the learning process has not yet fully achieved optimal effectiveness and still requires improvement. Centroid analysis shows that material understanding and learning interaction are the dominant factors influencing students’ learning difficulties. The clustering results provide a clear description of student characteristics within each group, which can support academic evaluation processes. Therefore, the findings of this study are expected to assist lecturers and study programs in developing more adaptive, effective, and data-driven learning strategies to improve the overall quality of higher education learning. Keywords: Data Mining, K-Means Clustering, Learning Difficulties, Students, Elbow Method