Clustering of Lecturer Performance Based on Internal Data and Questionnaire Using the X-Means Algorithm

Abstract

Lecturer performance evaluation is essential for improving the quality of higher education. In the Informatics Engineering Study Program at Universitas Muhammadiyah Jember, assessments are still conducted manually, leading to inefficiency and subjectivity. This study aims to cluster lecturer performance based on internal data—credit load, publications, attendance, and students’ average grades—and student perceptions collected through questionnaires, using the X-Means Clustering algorithm. X-Means was chosen because it can automatically determine the optimal number of clusters using the Bayesian Information Criterion (BIC) and log-likelihood. The research involved data collection, cleaning, normalization with the Min-Max method, implementation of X-Means, and visualization using Principal Component Analysis (PCA). The results identified five distinct clusters: high-performing lecturers, lecturers with high publication output, research-active lecturers with low student grades, low-performing lecturers, and stable-performing lecturers. These findings provide an objective view of lecturer performance variations and can serve as a basis for developing effective strategies for lecturer improvement and continuous performance evaluation