Change-point multivariable quantile regression to explore effect of weather variables on building energy consumption and estimate base temperature range

Abstract

Mean regression analysis may not capture associations that occur primarily in the tails of the outcome distribution. In this study, we focused on multiple weather factors to find the extent to which they impact heating-related gas consumption at higher quantiles. We used change-point multivariable quantile regression models to investigate distributional effects and heterogeneity in the gas consumption-related responses to weather factors. Subsequently, we analyzed quantile regression coefficients that corresponded to absolute differences in specific quantiles of gas consumption associated with a one-unit increase in weather factors. We found that the association of weather factors and gas consumption varied across 19 quantiles of gas consumption distribution. Heterogeneities varied between case study buildings: right tails of gas consumption for the community and educational buildings were more susceptible to weather factors than those of the healthcare building. The base temperature of the community buildings across quantiles of gas consumption indicated a flat trend, but the uncertainty ranges were relatively large compared with those for the community and educational buildings. The developed method in this study can be widely utilised to identify the most important factors and the extent to which they affect gas consumption at specific quantiles.