Title: A data envelopment analysis method with missing and zero values: ranking of universities and target setting with limited resources

Authors: Bibi Faheema Luckhoo; Arshad Ahmud Iqbal Peer

Addresses: Department of Applied Mathematical Sciences, University of Technology, Mauritius, La Tour Koenig, Pointe-aux-Sables, 11134, Mauritius ' Department of Applied Mathematical Sciences, University of Technology, Mauritius, La Tour Koenig, Pointe-aux-Sables, 11134, Mauritius

Abstract: Ranking of universities is now imperative due to globalisation of higher education. The three main rankings for universities differ by methodology and different results are observed. In this paper, a non-parametric approach, data envelopment analysis (DEA), is used for obtaining an aggregate ranking of universities based on the data of those main ranking systems. Since the data set contains zero and missing values, we propose a method where optimal rescaling is achieved in order to satisfy the positivity requirement of DEA while also preserving the original rankings. Due to the absence of input measures in our data set, we apply DEA models without explicit inputs (WEI) to the transformed data set in order to obtain a distinct ranking of the universities. We also propose a DEA method where targets can be set while decisions are made about allocation of resources in the context of WEI.

Keywords: DEA; data envelopment analysis; missing values; zero values; university rankings; target setting; limited resources; WEI; without explicit inputs; positivity requirement.

DOI: 10.1504/IJDS.2021.122771

International Journal of Data Science, 2021 Vol.6 No.4, pp.271 - 291

Received: 23 Feb 2021
Accepted: 20 Oct 2021
Published online: 10 May 2022 *

Full-text access for editors Full-text access for subscribers Purchase this article Comment on this article