| Article ID: | iaor201529079 |
| Volume: | 32 |
| Issue: | 6 |
| Start Page Number: | 707 |
| End Page Number: | 720 |
| Publication Date: | Nov 2015 |
| Journal: | Systems Research and Behavioral Science |
| Authors: | Fu Hsin-Pin, Chang Tien-Hsiang, Shieh Lon-Fon, Lin Arthur, Lin Shang-Wen |
| Keywords: | statistics: data envelopment analysis, neural networks, decision, retailing |
A better measurement tool can provide more accurate information to improve the evaluation of performance in terms of operational efficiency. The data envelopment analysis (DEA) method has been widely used for the measurement of performance in retail chain stores. However, if the data utilized in DEA are subject to statistical deviation, observational errors can occur in the measurement of operational efficiency, and the output can be distorted. Moreover, the explanatory power of DEA is relatively weak when analyzing multiple decision‐making units. To address these shortcomings, this paper used the back‐propagation neural network (BPN) in combination with DEA (DEA–BPN) to produce more reliable results. A leading chain of Taiwanese retail outlets selling lifestyle accessories was chosen as the sample to test the performance of the proposed DEA–BPN model. The results of the study verify that the DEA–BPN method is a more reliable tool to measure efficiency in a retail setting.