This release introduces significant changes to the internal logic behind the Malmquist index. The presentation of results is now consistent across the app, and the newly exported functions operate in a manner similar to other DEA functions.
Breaking changes
- Results for the Malmquist index are now returned ala Farrell (not Shephard). This is consistent with how results
for other models are currently presented in the app. Users should note that this changes the interpretation of the malmquist index for input-oriented models. See?compute_malmquist()for details. (#81) - Removed the select input box for returns to scale in the Malmquist analysis tab. The currently supported Malmquist index is the one defined by Färe & Grosskopf (1996). Components of the index is thus always computed based on the same returns to scale (either CRS or VRS depending on the specific component), so the select input box is not needed.
runPioneeR()has been removed to avoid confusion for new users with the release on CRAN.
Enhancements
- Added a new exported function,
compute_malmquist()for calculating productivity. This also removes the need for{productivity}as a dependency. (#81) - Further sub-components of the Malmquist index, including bias technical change and scale efficiency change, have been added to the results in the Malmquist analysis tab. (#81)
- Added caching to the bootstrap function in the app (#88)
- A warning is now generated if the power of the analysis is too low and more DMUs should be added for robust results (#83)
Full Changelog: v0.4.0...v0.5.0