SELECTING INPUTS AND OUTPUTS IN DATA ENVELOPMENT ANALYSIS BY DESIGNING STATISTICAL EXPERIMENTS( Operations Research for Performance Evaluation)

Abstract

Data envelopment analysis (DEA) is a data oriented, no"-parametric method to evaluate relative efficiency. based on pre-selected iiiputs aiid output･s. In some eases, t･he peTformance rnodel is not wel] defined, so it is critical to select the appropriate inputs and oui,puts by other means. When wc have many potential variables for evalua,t･ioii, iL is diMcult to select･ inputs and ouLputs ft/om a large number of possihle combinations. NNle propose an input output selection rnethod that uses diagonal layout･ experiments, which is a statistical approach to find an optimal combinaLion. We demonst,rate the proposed inethod using financial statement data from NIKKEI 500 index.