How Successful are Energy Efficiency Investments? A Comparative Analysis for Classification & Performance Prediction

Increasing the financial institutions’ deployment of capital in energy efficiency investments remains still a challenge. The present article is intended to investigate the benefits of the application of traditional classification methods, such as the ordinal logit, the ordinal probit and the linear discriminant analysis (LDA), as well as machine learning techniques, such as the k-Nearest Neighbors and the Support Vector Machines, in the development of models for predicting the performance of energy efficiency investments. We are dealing with the process of investments identification that can be considered attractive, in terms of fostering green growth, while also having an extremely strong capacity to meet their financial commitments and therefore bridging the gap between investors and project developers. In addition, the deduction of critical comparative insights regarding the application of these five widely used techniques, is anticipated. The validity of the attempt is verified through an illustrative testing procedure on the Energy Efficiency De-risking Project database. The qualitative and technical conclusions obtained demonstrate that machine learning methods moderately outperform traditional methods regarding their predictive accuracy. Finally, findings that confirm and expand the existing underlying research are also reported.