Data Envelopment Analysis with User-Defined and Learned Reference Sets

Slacks-based data envelopment analysis for panel data in which the reference set is an explicit argument. A user-supplied weight matrix, or a matrix learned from the data by penalised non-negative reconstruction, selects the units that may enter each decision-making unit's frontier. The package reports the unrestricted and the restricted score, the ratio of the two and an exact decomposition of the shortfall into the part the restricted technology can remove and the part it cannot. Moran's I and local indicators of spatial association are provided as diagnostics. All linear programs are solved by a two-phase simplex with Bland's rule implemented in R, so the package needs no compiler.

psdea

Reference-set data envelopment analysis for panel data.

An efficiency score is a distance to a frontier, and the frontier is built from peers. psdea makes the peer set an argument: a weight matrix, supplied by the user or learned from the data, selects which units may enter each unit's frontier. The package returns the unrestricted and the restricted score, their ratio, and an exact split of the shortfall.

Install

R CMD INSTALL psdea_1.0.0.tar.gz

Use

library(psdea)
data(psdea_china_energy); data(psdea_w_contiguity)
d <- subset(psdea_china_energy, year <= 2022)
fit <- psdea_panel(d, "gdp_real", c("capital", "energy"), "year",
                   id = "code", bad = "co2", W = psdea_w_contiguity)
fit

Design

• Only the pattern of the matrix is used; weights are read as indicators.
• The unit itself is always in its own reference set.
• Shares are NA on the frontier rather than zero.
• No compiled code, no dependencies outside base R.