Data envelopment analysis

Summary

The software makes a data envelopment analysis (DEA) and an alternative DEA, called ADEA, to redistribute the loads between variables. Futhermore, using the loads of the variables, it makes an hierarchical analysis to select the most outstanding variables.

Use

Select one action in, complete the values and click the "Go" button.

Steps:

1. As first step load an example, load a data set or build a new one.
2. Select "Model" tab.
3. Now, if you want, you can change the orientation of the model.
4. Now add inputs and outputs to the model. You must select at least one input and one output.
5. Select the tab with the results you want to show.

You can enter new values and the results will be updated automatically.

Model selection

To select the variables in the model, as first step add to model the variables for the initial one selection "Add input" and "Add output" as action. After that select the action "Model selection" in the left panel and set the number variables or the load value, or both, to the desired values. Then press "Go" botton and the results for the new model will be shown.

Outputs

In "Graphic" tab all the pairs of variables in data set are plotted.

"Model" tab shows the name of the variables in the data set, the orientation for the AED analysis and the rol in it of each variable. Notice that only numeric variables can be included in AED.

"DEA Results" tab show the results of the standar data envelopment analysis. The name of input variables is show preceded by i_. Anagously the name of output variables are preceded by o_. For both the weights in the analysis are shown. For the other variables its values is what is shown. One can search into the results and reorder by columns.

"ADEA Results" works identicaly as the previous one but doing the alternative analysis.

The tabs "Hierarchical Model" and "Parametric Model" requires at least one more input or output. In these tabs the result of stepwise selection of the variables in the model are shown.

References

• Stepwise selection of variables in DEA using contribution loads. Fernando Fernandez-Palacin, Maria Auxiliadora Lopez-Sanchez, Manuel Munoz-Marquez. Pesquisa Operacional, 38:1, pg. 31-52. 2018, DOI: 10.1590/0101-7438.2018.038.01.0031.

Note: In some cases the computations require a long time, please, have patience.

Developed using shiny by M. Muñoz-Márquez under GNU-GPL licence.

There is a forum available to make question and suggestions, and more information.