Abstract
In nonparametric methods, if the number of observations is
relatively small as compared to the sum of number of inputs and
outputs, many
units are evaluated as efficient. Several methods
for prioritizing these efficient units are reported in literature.
Andersen et al. and Mehrabian et al.
proposed two methods for ranking efficient
units, but both methods break down in some cases. This paper
describes a new DEA ranking approach that
uses