This PR reworks the compute_efficiency() and compute_dea() functions such that a revised S3 object is returned from the main exported function. This object can in return be used with the new bootstrap_dea() function. This function enables the user to estimate confidence intervals and bias corrected efficiency scores based on Simar & Wilson, 1998. This PR should include all functionality to run the bootstrap in an R session and within the app. A very simple R script to estimate bias corrected scores for the Färe data set is as follows:
# Get data
fare89 <- deaR::Electric_plants
# Estimate efficiency
mod <- compute_dea(fare89, 'Plant', c('Labor', 'Fuel', 'Capital'), 'Output', 'vrs', 'in')
# Run bootstrap
boot <- bootstrap_dea(mod, iterations = 2000)
# Using parameters in Simar & Wilson
boot_sw <- bootstrap_dea(mod, bw_rule = 0.014, iterations = 2000)
This is still considered a draft. I will work on examples and tests, and add these in a separate commit.
This PR reworks the
compute_efficiency()
andcompute_dea()
functions such that a revised S3 object is returned from the main exported function. This object can in return be used with the newbootstrap_dea()
function. This function enables the user to estimate confidence intervals and bias corrected efficiency scores based on Simar & Wilson, 1998. This PR should include all functionality to run the bootstrap in an R session and within the app. A very simple R script to estimate bias corrected scores for the Färe data set is as follows:This is still considered a draft. I will work on examples and tests, and add these in a separate commit.
This PR will also close #73 and #89.