Variable Importance Analysis for Urban Building Energy Assessment in the Presence of Correlated Factors

Song Yang, Wei Tian, Yeonsook Heo, Qingxing Meng, Lai Wei

Research output: Contribution to journalConference articlepeer-review

4 Citations (Scopus)

Abstract

It is becoming urgent to thoroughly understand characteristics of energy use in order to reduce energy use in urban areas. When assessing energy performance in urban buildings, it is likely that explanatory variables are correlated if considering both physical conditions and social economic factors. This research applied three variable importance methods, including Genizi, CAR (Correlation-Adjusted marginal coRrelation), PCC (partial correlation coefficient), to identify key factors from 30 highly correlated variables in London. The results indicate that the land area for domestic buildings is the only dominant variable influencing gas use, while electricity consumption is more affected by the number of electricity meters for Economy 7 (a differential electricity tariff according to the time of day) and the number of households allocated to higher council tax band in London. Moreover, it is confirmed that the SRC (standardized regression coefficient), a commonly used method in building energy analysis, is not suitable for the correlated factors in urban energy assessment.

Original languageEnglish
Pages (from-to)277-284
Number of pages8
JournalProcedia Engineering
Volume121
DOIs
Publication statusPublished - 2015
Externally publishedYes
Event9th International Symposium on Heating, Ventilation and Air Conditioning, ISHVAC 2015 Joint with the 3rd International Conference on Building Energy and Environment, COBEE 2015 - Tianjin, China
Duration: 2015 Jul 122015 Jul 15

Keywords

  • Building energy
  • Correlated variables
  • Urban environment
  • Variable importance

ASJC Scopus subject areas

  • Engineering(all)

Fingerprint

Dive into the research topics of 'Variable Importance Analysis for Urban Building Energy Assessment in the Presence of Correlated Factors'. Together they form a unique fingerprint.

Cite this