A new approach based on data envelopment analysis with double frontiers for ranking the discovered rules from data mining

Data envelopment analysis (DEA) is a relatively new data oriented approach to evaluate performance of a set of peer entities called decision-making units (DMUs) that convert multiple inputs into multiple outputs. Within a relative limited period, DEA has been converted into a strong quantitative and analytical tool to measure and evaluate performance. In an article written by Toloo et al. (2009), they proposed a new DEA model to find the most efficient association rule in data mining. Considering several criteria, they created an algorithm for ranking association rules using this model. In the present article, we show that their model only selects an optimistic efficient association rule randomly and it is completely dependent on solution or software, which is used for solving problems. In addition, it shows that their proposed algorithm can only rank optimistic efficient rules randomly and it is not able to rank optimistic non-efficient DMUs. We mention other disadvantages and propose a new approach “DEA with double frontiers” to create a complete ranking of association rules. A numerical example will explain some contents of the paper.