Analysis of Academic Staff Performance With Data Mining

Abstract

This study focuses on analyzing the performance of academics at a university in Istanbul based on six main criteria aligned with the university's performance evaluation guidelines. These criteria include publications (articles, books, and conference papers), teaching load, supervision of theses, and administrative responsibilities. A comprehensive dataset was compiled using data sourced from the Higher Education Council (YÖK) and university records pertaining to these criteria. The dataset, which initially consisted of 12664 entries representing 199 academics from nine different research domains, is transformed into input data containing 1194 entries. Subsequently, the dataset is clustered into five groups via K-means algorithm utilizing WEKA software. The distinct performance characteristics of the clusters are identified and discussed. The genders and research domains of the academic staff are also considered in the final discussions. The findings were presented to the University Administration in order to provide insight into strategic decisions.