DATA ENVELOPMENT ANALYSIS WITH MISSING DATA: A MULTIPLE LINEAR REGRESSION ANALYSIS APPROACH

Abstract

Data envelopment analysis (DEA) assumes that the data set is precise when performing efficiency evaluation of peer decision making units (DMUs). The current paper proposes a multiple linear regression analysis (MLRA) approach to estimate missing values if some of the entries in the data set are missing. Its algorithm to derive the estimations is also proposed. In order to verify the credibility of the proposed approach, an example of 30 US commercial banks is applied to case analysis. Using the proposed algorithm, the efficiencies of all DMUs are obtained. A Friedman test and a Kendall's Tau rank correlation analysis statistically examine the results. Moreover, the efficiency interval and efficiency distribution for a DMU are obtained considering random errors of the estimations. After that, an example of public secondary schools serves to illustrate the applications in the end.