Mining Scientific Results Through the Combined Use of Clustering and Linear Programming Techniques

The paper proposes a technique based on a combined approach of data mining algorithms and linear programming methods for classifying organizational units, such as research centers. We exploit clustering algorithms for grouping information concerning the scientific activity of research centers. We also show that the replacement of an expensive efficiency measurement, based on the solution of linear programs, with a simple formula allows clusters of very good quality to be computed efficiently. Some initial experimental results, obtained from an analysis of research centers in the agro-food sector, show the effectiveness of our approach, both from an efficiency and a quality-of-results point of view.