Comparing Approaches to Handling Data Envelopment Analysis with Missing Values using the Case of Machinery Industry in Taiwan

Research areas:

• Volume 3,Issue 6, May 2015

Year:

2015

Type of Publication:

Article

Keywords:

Data Envelopment Analysis, Fuzzy, Machinery Industry, Missing Values

Authors:

Hsu-Hao Yang; Chia-Jung Kao

Journal:

IJAIM

Volume:

3

Number:

6

Pages:

267-274

Month:

May

Abstract:

The purpose of this paper is to investigate existing methods for dealing with missing values in a data envelopment analysis (DEA) model. Three types of methods are investigated: (1) fuzzy DEA (FDEA), (2) interval DEA (IDEA), and (3) the method that codes missing values with appropriate values. In essence, the FDEA and method (3) calculate and rank efficiency estimates, whereas the IDEA calculates and classifies estimates into categories. We apply the three methods to the data of 17 Taiwanese companies making machine tools and conclude as follows. First, both the FDEA and method (3) produce consistent rankings of estimates. Next, companies performing relatively well are identified by each method. In addition, companies with missing values appear to have identical relative sequence of estimates regardless of the method used. Moreover, the presence of missing values could have little effect on the ranking. Finally, companies classified by the IDEA may rank discrepantly by the FDEA, which suggests that both the FDEA and IDEA can be used together to provide complementary insights for making decisions involving trade-offs.

Full text: IJAIM_438_Final.pdf