期刊名称：International Journal of Computer Science and Network Security
印刷版ISSN：1738-7906
出版年度：2017
卷号：17
期号：2
页码：74-82
出版社：International Journal of Computer Science and Network Security
摘要：Data envelopment analysis (DEA) is a popular approach for measuring the relative efficiency of homogenous units that utilize multiple inputs to produce multiple outputs. In spite of few researches on the relationship between clustering approach and DEA, this paper proposes an in-depth look at conceptual definition of the performance of clustering units. This study is different in a very significant way specifically two kinds of approaches were integrated to develop the model. The first method is context-dependent DEA proposed by Seiford et al. (2003) which have formed the basis of many previous studies. The second method is obtained from finding degree-DMU, since finding degree-unit is always a concern. Andersen et al. (1993) have proposed a model for finding super-efficient DMU. The main reason for applying the super efficiency approach is that: (i) in a group of people consisting of president (CEO), the vice president, the manager and the general public, it is a rational way of putting each specific member in its relevant cluster, (ii) for each cluster, a cluster ranking orders the members, (iii) if we number the clusters from 1 up to r, then cluster 1istop priority, cluster 2 has the second highest priority, etc. This paper is intended to cluster all DMUs with the help of these two approaches. Additionally, we compared our approach with context dependent DEA, and finally, the proposed approach has been applied to classify 25 branches of an Iranian commercial bank.
关键词：Data envelopment analysis clustering Degree-DMU context-dependent decision making units.
