Pursuing phased improvement in organizational performance: A hybrid system

This paper proposes a hybrid data envelopment analysis (DEA) system that utilizes a methodology combining the tier analysis with the neural clustering method. It aims to show that the hybrid system can be used to evaluate the inter-organizational efficiency in the life insurance companies. The application is unfolded in two phases. In the first phase, DEA is repetitively used to evaluate the efficiency of DMUs and cluster them together according to their efficiency levels (tier analysis). In the second phase, the system utilizes a self organizing map to group similar DMUs, selects benchmarking targets within a reference set, and provides the guidelines on the stepwise enhancements for the inefficient ones.