[R] DEA question

John P. Burkett burkett at uri.edu
Wed Sep 24 07:07:45 CEST 2008


Trying to learn to use the DEA package in R version 2.7.1 running under 
Gentoo  Linux, I have applied dea.ccr.io.mul() and dea.ccr.oo.mul() to 
very simple artificial data. dea.ccr.io.mul() has produced exactly the 
results I expected.  In contrast, dea.ccr.oo.mul() produces results that 
surprise me.

My code for second case is as follows:
# This is a file of R commands to illustrate output-oriented
# data envelopment analysis.
# Step 1:  Enter input and output data
X <- rbind(1,
            1,
            1)
Y <- rbind(c(1, 3),
            c(3, 1),
            c(1, 1))
# Step 2: load the DEA package and use it to analyze the data
library(DEA)
results <- dea.ccr.oo.mul(X, Y, infor=TRUE)

My expectation was that the results would indicate that DMU1 and DMU2 
have efficiency 1 and DMU3 has efficiency 0.5.  Instead the results are 
as follows:

$eff
DMU1 DMU2 DMU3
    1    1    2

$weights
      v.x1      u.y1      u.y2
DMU1    1 0.0000000 0.3333333
DMU2    1 0.3333333 0.0000000
DMU3    2 0.5000000 0.5000000

It appears to me that the reported "eff" for DMU3 is the reciprocal of 
efficiency.  Is this a bug or am I misunderstanding something?

Best regards,
John

-- 
John P. Burkett
Department of Environmental and Natural Resource Economics
and Department of Economics
University of Rhode Island
Kingston, RI 02881-0808
USA

phone (401) 874-9195



More information about the R-help mailing list