An AIC-based Approach to Identify the Most Influential Variables in Eco-efficiency Evaluation

Abstract Eco-efficiency evaluation has received increasing public attention and plays an important role in the business community. In many practical applications, the decision-makers are interested in which eco-variables take a significant effect on eco-efficiency evaluation and how to select proper variables in situations where there are a large number of alternative variables. This paper approaches these problems based upon the Akaike information criteria (AIC) rule. The proposed approach can investigate all possible variable sets and identify the most influential variables. A real data set about 30 industrial systems in China has been used to illustrate the proposed approach. We find the most influential undesirable output in determining provincial industrial systems' eco-efficiency of China is Sulphur dioxide emission. This result is robust under different eco-efficiency measurements. It is of great significance for decision-makers to achieve eco-efficiency improvement.