Table 2

Summary of literature on clustering analysis used in combination with DEA

AuthorsContext of analysisClustering methodClustering variablesNr. of DMUsNr of clustersMethod of combination with DEA
Samoilenco and Esei-Bryson (2008, 2010) Countries in transition from centralized economies to market economiesTwo-step approach based on k-meansLevels of DMUs inputs and outputs182 clusters, through user-specified threshold and domain expert knowledgeClustering is adopted to DEA results on the entire dataset
Sharma and Yu (2009) Container terminalsKohonen's Self Organizing Map (KSOM) preceded by a stratifying methodLevels of DMUs inputs704 clusters through unsupervised clusteringClustering is adopted to DEA results on the entire dataset
Amirteimoori and Kordrostami (2013) Retail: Bank branchesOriginal method based on branches sizeSize643 clusters by domain expert knowledgeClustering is adopted prior to DEA. DEA is used within clusters
Hajiagha et al. (2016) Retail: Bank branchesFuzzy c-means clusteringLevels of DMUs inputs and outputs1172 clusters set by the analystClustering is adopted prior to DEA. DEA is used within clusters
Li et al. (2016) Retail: Gas stationsWard's hierarchical clusteringLevels of DMUs inputs and outputs1974 clusters set by the analystClustering is adopted to DEA results on the entire dataset
Omrani et al. (2018) HospitalsFuzzy c-means clusteringEnvironmental characteristics (population; GDP per capita)2885 clusters set by the analystClustering is adopted prior to DEA. DEA is used within clusters
Costa et al. (2019) Electricity energy distribution utilitiesSpatial Bayesian clusteringSpatial location (assuming that geographically closer DMUs are homogeneous)642 clusters as result of the analysisClustering is adopted to DEA results on the entire dataset
Samoilenko and Osei-Bryson (2019) Sub-Saharan African countriesHybrid partitional/hierarchical approachEconomic development; socioeconomic impact of ICT; growth in productivity273 clusters by domain expert knowledgeClustering is adopted to DEA results on the entire dataset
Cinaroglu (2020) HospitalsK-means clusteringregional areas are clustered on welfare indicators815 clusters through a combination of different factorsClustering is adopted prior to DEA. DEA is used within clusters
Zarrin et al. (2022) HospitalsSelf-Organizing Map-Artificial Neural Networkhospital's characteristics1,1243 clusters through unsupervised clusteringClustering is adopted prior to DEA. DEA is used within clusters
Tsionas (2023) Commercial banksConvex non-parametric least squaresCommercial banks operating variables nd technical feasibility2853 clusters through Bayesian Model AveragingDEA is used on the entire dataset. Clustering is adopted to DEA results
Our contributionRetail: Kitchen furnitureFactorial techniques (PCA and MCA) on Big Data + Combined Agglomerative Hierarchical Clustering and k-means algorithm methodStructural characteristics of stores and socio-demographic characteristics of their catchment area5413 clusters as result of an agglomerative hierarchical clustering, based on factorial scoresClustering is adopted prior to DEA. DEA is used within clusters

or Create an Account

Close Modal
Close Modal