Tuning machine learning models for prediction of building energy loads

Saleh Seyedzadeh, Farzad Pour Rahimian Leilabadi, Parag Rastogi, Ivan Glesk

Research output: Contribution to journalArticle

2 Citations (Scopus)

Abstract

There have been numerous simulation tools utilised for calculating building energy loads for efficient design and retrofitting. However, these tools entail a great deal of computational cost and prior knowledge to work with. Machine Learning (ML) techniques can contribute to bridging this gap by taking advantage of existing historical data for forecasting new samples and lead to informed decisions. This study investigated the accuracy of most popular ML models in the prediction of buildings heating and cooling loads carrying out specific tuning for each ML model and using two simulated building energy data generated in EnergyPlus and Ecotect and compared the results. The study used a grid-search coupled with cross-validation method to examine the combinations of model parameters. Furthermore, sensitivity analysis techniques were used to evaluate the importance of input variables on the performance of ML models. The accuracy and time complexity of models in predicting heating and cooling loads are demonstrated. Comparing the accuracy of the tuned models with the original research works reveals the significant role of model optimisation. The outcomes of the sensitivity analysis are demonstrated as relative importance which resulted in the identification of unimportant variables and faster model fitting.

LanguageEnglish
Article number101484
Number of pages18
JournalSustainable Cities and Society
Volume47
Early online date16 Mar 2019
DOIs
Publication statusPublished - 31 May 2019

Fingerprint

Dynamic loads
Learning systems
building
Tuning
energy
prediction
learning
heat pump
grid search
Sensitivity analysis
sensitivity analysis
Cooling
heating
cooling
Heating
optimization model
machine learning
Retrofitting
research work
Loads (forces)

Keywords

  • building energy loads
  • energy prediction
  • machine learning
  • energy modelling
  • energy simulation
  • building design

Cite this

@article{a696cbc7d8b444feadb4299b6ec32992,
title = "Tuning machine learning models for prediction of building energy loads",
abstract = "There have been numerous simulation tools utilised for calculating building energy loads for efficient design and retrofitting. However, these tools entail a great deal of computational cost and prior knowledge to work with. Machine Learning (ML) techniques can contribute to bridging this gap by taking advantage of existing historical data for forecasting new samples and lead to informed decisions. This study investigated the accuracy of most popular ML models in the prediction of buildings heating and cooling loads carrying out specific tuning for each ML model and using two simulated building energy data generated in EnergyPlus and Ecotect and compared the results. The study used a grid-search coupled with cross-validation method to examine the combinations of model parameters. Furthermore, sensitivity analysis techniques were used to evaluate the importance of input variables on the performance of ML models. The accuracy and time complexity of models in predicting heating and cooling loads are demonstrated. Comparing the accuracy of the tuned models with the original research works reveals the significant role of model optimisation. The outcomes of the sensitivity analysis are demonstrated as relative importance which resulted in the identification of unimportant variables and faster model fitting.",
keywords = "building energy loads, energy prediction, machine learning, energy modelling, energy simulation, building design",
author = "Saleh Seyedzadeh and {Pour Rahimian Leilabadi}, Farzad and Parag Rastogi and Ivan Glesk",
year = "2019",
month = "5",
day = "31",
doi = "10.1016/j.scs.2019.101484",
language = "English",
volume = "47",
journal = "Sustainable Cities and Society",
issn = "2210-6707",

}

Tuning machine learning models for prediction of building energy loads. / Seyedzadeh, Saleh; Pour Rahimian Leilabadi, Farzad; Rastogi, Parag; Glesk, Ivan.

In: Sustainable Cities and Society, Vol. 47, 101484, 31.05.2019.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Tuning machine learning models for prediction of building energy loads

AU - Seyedzadeh, Saleh

AU - Pour Rahimian Leilabadi, Farzad

AU - Rastogi, Parag

AU - Glesk, Ivan

PY - 2019/5/31

Y1 - 2019/5/31

N2 - There have been numerous simulation tools utilised for calculating building energy loads for efficient design and retrofitting. However, these tools entail a great deal of computational cost and prior knowledge to work with. Machine Learning (ML) techniques can contribute to bridging this gap by taking advantage of existing historical data for forecasting new samples and lead to informed decisions. This study investigated the accuracy of most popular ML models in the prediction of buildings heating and cooling loads carrying out specific tuning for each ML model and using two simulated building energy data generated in EnergyPlus and Ecotect and compared the results. The study used a grid-search coupled with cross-validation method to examine the combinations of model parameters. Furthermore, sensitivity analysis techniques were used to evaluate the importance of input variables on the performance of ML models. The accuracy and time complexity of models in predicting heating and cooling loads are demonstrated. Comparing the accuracy of the tuned models with the original research works reveals the significant role of model optimisation. The outcomes of the sensitivity analysis are demonstrated as relative importance which resulted in the identification of unimportant variables and faster model fitting.

AB - There have been numerous simulation tools utilised for calculating building energy loads for efficient design and retrofitting. However, these tools entail a great deal of computational cost and prior knowledge to work with. Machine Learning (ML) techniques can contribute to bridging this gap by taking advantage of existing historical data for forecasting new samples and lead to informed decisions. This study investigated the accuracy of most popular ML models in the prediction of buildings heating and cooling loads carrying out specific tuning for each ML model and using two simulated building energy data generated in EnergyPlus and Ecotect and compared the results. The study used a grid-search coupled with cross-validation method to examine the combinations of model parameters. Furthermore, sensitivity analysis techniques were used to evaluate the importance of input variables on the performance of ML models. The accuracy and time complexity of models in predicting heating and cooling loads are demonstrated. Comparing the accuracy of the tuned models with the original research works reveals the significant role of model optimisation. The outcomes of the sensitivity analysis are demonstrated as relative importance which resulted in the identification of unimportant variables and faster model fitting.

KW - building energy loads

KW - energy prediction

KW - machine learning

KW - energy modelling

KW - energy simulation

KW - building design

UR - https://www.sciencedirect.com/journal/sustainable-cities-and-society

U2 - 10.1016/j.scs.2019.101484

DO - 10.1016/j.scs.2019.101484

M3 - Article

VL - 47

JO - Sustainable Cities and Society

T2 - Sustainable Cities and Society

JF - Sustainable Cities and Society

SN - 2210-6707

M1 - 101484

ER -