Development of a Decision Support System for Clustering Scientific Publications Using K-Means

Abstract

To improve the quality of departments within a university, a decision support system ( DSS ) that maps the number of scientific publications published is needed urgently today. This is because clusters and ranks for a university are determined by these scientific publications. In DSS , there is a crawler bot to recap publication data from Google Scholar and Scopus, as well as the K-means algorithm for clustering departments. DSS will produce 2 clusters with a silhouette score of 0.6245660527900866. A positive value indicates that the data in one cluster are similar. The sample used data from faculty members at the Faculty of Teacher Training and Education at Sebelas Maret University in Indonesia. The results showed that cluster one consisted of 18 departments. The average results were Scopus publication of 0.4669, Google scholar publications of 1.8955, Scopus citations of 0.8328, Google scholar citations of 18.3519. Cluster two consists of six departments., The average results were Scopus publication of 3.1727, Google Scholar publication of 4.2636, Scopus Citation of 9.0636, Google Scholar Citation of 67.6455.