A Theorem on Weak DEA Efficiency (C~2GS~2) and Its Application in Forecast

摘要

The paper gives a theorem on making a decision making unit (DMU) weakly DEA efficient (C~2GS~2) by changing its inputs only or its outputs only. The theorem is used to forecast a problem with multiple inputs and multiple outputs. To make the process of forecasting with new method more clear, we take some production activities with single input and single output as an example. Suppose that there are production activities (X_j, Y_j) (j=1, ..., n), from which we get a production possibility set. The effective frontier of this set is used for forecasting, i.e., for a given input X_(n+1), if (X_(n+1),Y_(n+1)) lies on the effective frontier, then Y_(n+1) is taken as the predicted value of output, and it is the maximal value in all possible output for the given input X_(n+1) . While for the given X_(n+1), the predicted value of output obtained from the regression analysis on the given production activities is the mean value of all possible outputs.