Evaluation of tree-based ensemble learning algorithms for building energy performance estimation

Sokratis Papadopoulos, Elie Azar, Wei Lee Woon, Constantine Kontokosta

Research output: Contribution to journalArticle

Abstract

Tree-based ensemble learning has received significant interest as one of the most reliable and broadly applicable classes of machine learning techniques. However, thus far, it has rarely been used to model and evaluate the drivers of energy consumption in buildings and as such its performance and accuracy in this field have yet to be properly tested or fully understood. The goal of this paper is to evaluate the performance of three ensemble learning algorithms in modelling and predicting the heating and cooling loads of buildings, namely (i) random forests, (ii) extremely randomized trees (extra-trees), and (iii) gradient boosted regression trees. Results show that the tested algorithms outperform the ones proposed in the recent literature, with gradient boosting improving on the prediction accuracy of the second best-performing algorithm by an average of 14% and 65% for the heating and cooling loads, respectively.

Original languageEnglish (US)
Pages (from-to)1-11
Number of pages11
JournalJournal of Building Performance Simulation
DOIs
StateAccepted/In press - Jul 28 2017

Fingerprint

Ensemble Learning
Learning algorithms
Cooling
Heating
Learning Algorithm
Gradient
Regression Tree
Random Forest
Evaluate
Evaluation
Boosting
Energy
Energy Consumption
Driver
Learning systems
Loads (forces)
Machine Learning
Energy utilization
Prediction
Modeling

Keywords

  • artificial intelligence
  • building energy performance
  • ensemble learning
  • extra-trees
  • gradient boosting
  • heating ventilation and air conditioning
  • random forests

ASJC Scopus subject areas

  • Architecture
  • Modeling and Simulation
  • Building and Construction
  • Computer Science Applications

Cite this

Evaluation of tree-based ensemble learning algorithms for building energy performance estimation. / Papadopoulos, Sokratis; Azar, Elie; Woon, Wei Lee; Kontokosta, Constantine.

In: Journal of Building Performance Simulation, 28.07.2017, p. 1-11.

Research output: Contribution to journalArticle

@article{5edd05d229f04b589b6e73834830764c,
title = "Evaluation of tree-based ensemble learning algorithms for building energy performance estimation",
abstract = "Tree-based ensemble learning has received significant interest as one of the most reliable and broadly applicable classes of machine learning techniques. However, thus far, it has rarely been used to model and evaluate the drivers of energy consumption in buildings and as such its performance and accuracy in this field have yet to be properly tested or fully understood. The goal of this paper is to evaluate the performance of three ensemble learning algorithms in modelling and predicting the heating and cooling loads of buildings, namely (i) random forests, (ii) extremely randomized trees (extra-trees), and (iii) gradient boosted regression trees. Results show that the tested algorithms outperform the ones proposed in the recent literature, with gradient boosting improving on the prediction accuracy of the second best-performing algorithm by an average of 14{\%} and 65{\%} for the heating and cooling loads, respectively.",
keywords = "artificial intelligence, building energy performance, ensemble learning, extra-trees, gradient boosting, heating ventilation and air conditioning, random forests",
author = "Sokratis Papadopoulos and Elie Azar and Woon, {Wei Lee} and Constantine Kontokosta",
year = "2017",
month = "7",
day = "28",
doi = "10.1080/19401493.2017.1354919",
language = "English (US)",
pages = "1--11",
journal = "Journal of Building Performance Simulation",
issn = "1940-1493",
publisher = "Taylor and Francis Ltd.",

}

TY - JOUR

T1 - Evaluation of tree-based ensemble learning algorithms for building energy performance estimation

AU - Papadopoulos, Sokratis

AU - Azar, Elie

AU - Woon, Wei Lee

AU - Kontokosta, Constantine

PY - 2017/7/28

Y1 - 2017/7/28

N2 - Tree-based ensemble learning has received significant interest as one of the most reliable and broadly applicable classes of machine learning techniques. However, thus far, it has rarely been used to model and evaluate the drivers of energy consumption in buildings and as such its performance and accuracy in this field have yet to be properly tested or fully understood. The goal of this paper is to evaluate the performance of three ensemble learning algorithms in modelling and predicting the heating and cooling loads of buildings, namely (i) random forests, (ii) extremely randomized trees (extra-trees), and (iii) gradient boosted regression trees. Results show that the tested algorithms outperform the ones proposed in the recent literature, with gradient boosting improving on the prediction accuracy of the second best-performing algorithm by an average of 14% and 65% for the heating and cooling loads, respectively.

AB - Tree-based ensemble learning has received significant interest as one of the most reliable and broadly applicable classes of machine learning techniques. However, thus far, it has rarely been used to model and evaluate the drivers of energy consumption in buildings and as such its performance and accuracy in this field have yet to be properly tested or fully understood. The goal of this paper is to evaluate the performance of three ensemble learning algorithms in modelling and predicting the heating and cooling loads of buildings, namely (i) random forests, (ii) extremely randomized trees (extra-trees), and (iii) gradient boosted regression trees. Results show that the tested algorithms outperform the ones proposed in the recent literature, with gradient boosting improving on the prediction accuracy of the second best-performing algorithm by an average of 14% and 65% for the heating and cooling loads, respectively.

KW - artificial intelligence

KW - building energy performance

KW - ensemble learning

KW - extra-trees

KW - gradient boosting

KW - heating ventilation and air conditioning

KW - random forests

UR - http://www.scopus.com/inward/record.url?scp=85026387190&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85026387190&partnerID=8YFLogxK

U2 - 10.1080/19401493.2017.1354919

DO - 10.1080/19401493.2017.1354919

M3 - Article

AN - SCOPUS:85026387190

SP - 1

EP - 11

JO - Journal of Building Performance Simulation

JF - Journal of Building Performance Simulation

SN - 1940-1493

ER -