Segmentation of Lecturer Productivity from Indexed Publications with K-Means Algorithm and Principal Component Analysis

Lecturers have an obligation to conduct research as a form of contribution to knowledge. Scientific publications, especially those indexed by Scopus, Google Scholar, and Web of Science (WoS), have a significant impact on the reputation of lecturers and the quality of universities. Analyzing the productivity of lecturers in conducting research and indexed publications is necessary to understand lecturers' productivity. Segmenting lecturer productivity based on indexed publications serves as an evaluation of lecturers' performance and enhances scientific publications. This research applies the K - Means Algorithm for clustering indexed publications of lecturers. Several productivity metrics used are based on indexed publications in Scopus, Google Scholar, WoS, and SINTA score. Additionally, attention is given to indicators such as h-index, i10-index, and g-index as references for the impact of publications used as publication references for other researchers. The segmentation results from various metrics are presented using Principal Component Analysis (PCA) approach to reduce the number of dimensions used in segmentation analysis. Additionally, information is obtained that generally, the number of Scopus-indexed publications is higher in Bachelor's degree programs, and Master's degree programs generally contribute more as publication impact. Based on experimental results with <tex xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">$k=4$</tex> from 145 faculty members, the best, moderate, and lower clusters were obtained. The best cluster has 6 lecturers with high productivity in Scopus publications. The total number of Scopus-indexed publications reaches 506 articles and cited by other publications reaches 2186 times with a SINTA score reaching 17140.