Selecting the outsourcing activities through the hybrid model of feature selection based on Kohonen network and slack-based measure

Abstract

One of the most important organisational, strategic decisions in the field of science and practice is the problem of outsourcing decision on business activities. In this problem, decision would turn various quantitative and qualitative criteria of outsourcing decision naturally into a complicated decision problem. If the intention of the decision-maker were focused on a portfolio of outsourcing decisions on business activities instead of selecting or deselecting a qualified activity for outsourcing, the study presented in the current paper could be applied. Based on the hybrid three-step methodology proposed in the current paper, the qualitative and quantitative criteria influencing outsourcing decision should be first identified through the feature selection algorithm to select the outsourcing business activities successfully. The advantage of such action would limit the large dimensions of the decision problem. Considering the identified features in the previous step, the qualified activities would be clustered for outsourcing the business activities through Kohonen neural network in the second step. Clustering the outsourced activities would integrate the qualified activities for outsourcing. In the final step, the efficiency of each cluster would be separately measured through DEA-SBM model. The methodology proposed in the current paper was applied through a case study in CNG stations.