Comparison of domestic lifestyle energy consumption clustering approaches: Fid-Bau Portal

Comparison of domestic lifestyle energy consumption clustering approaches

Bogin, Diana / Kissinger, Meidad / Erell, Evyatar

Highlights • Segmentation by hierarchical cluster analysis can be performed on small samples. • Clustering criteria affect outcome and must be appropriate for the study. • Classification by household attributes outperformed classifying by behavior. • Classification by behavior identified small high-consumption households.

Abstract The need to promote energy conservation measures in residential buildings has been widely acknowledged. However, policies supporting such measures may be inadequate unless they recognize and address the effect of individuals and households lifestyle on energy consumption. This study compares two clustering approaches to segmenting a population by lifestyle using hierarchical cluster analysis: electricity use behavior (EUB) and household characteristics (HCH). The method was demonstrated in the dry Mediterranean city of Beer Sheva, Israel. The HCH approach appeared better than EUB at classifying households into distinct sub-groups of similar size, with significantly distinct annual electricity consumption. The results demonstrate the importance of selecting appropriate criteria for analysis, as they impact the identification of clusters and the policies that could be drafted based upon such research in the general case.