Article ID | Journal | Published Year | Pages | File Type |
---|---|---|---|---|
729957 | Measurement | 2014 | 7 Pages |
•We develop a modified SORM model in dealing with negative data.•The modified SORM model can deal with input/output variables with both positive and negative values.•The modified SORM model is always feasible and bounded.•The modified SORM overcomes issues of target settings in the SORM.
Over the last few years Data Envelopment Analysis (DEA) has been gaining increasing popularity as a tool for measuring efficiency and productivity of Decision Making Units (DMUs). Conventional DEA models assume non-negative inputs and outputs. However, in many real applications, some inputs and/or outputs can take negative values. Recently, Emrouznejad et al. [6] introduced a Semi-Oriented Radial Measure (SORM) for modelling DEA with negative data. This paper points out some issues in target setting with SORM models and introduces a modified SORM approach. An empirical study in bank sector demonstrates the applicability of the proposed model.