Please use this identifier to cite or link to this item: https://doi.org/10.1080/23744731.2022.2067466
Title: Limitations of machine learning for building energy prediction: ASHRAE Great Energy Predictor III Kaggle competition error analysis
Authors: MILLER, CLAYTON 
PICCHETTI, BIANCA
FU, CHUN
PANTELIC, JOVAN
Keywords: Science & Technology
Physical Sciences
Technology
Thermodynamics
Construction & Building Technology
Engineering, Mechanical
Engineering
AUTOMATED MEASUREMENT
CONSUMPTION
VERIFICATION
UNCERTAINTY
LOAD
SAVINGS
MODELS
METHODOLOGY
PERFORMANCE
ACCURACY
Issue Date: 3-May-2022
Publisher: TAYLOR & FRANCIS INC
Citation: MILLER, CLAYTON, PICCHETTI, BIANCA, FU, CHUN, PANTELIC, JOVAN (2022-05-03). Limitations of machine learning for building energy prediction: ASHRAE Great Energy Predictor III Kaggle competition error analysis. SCIENCE AND TECHNOLOGY FOR THE BUILT ENVIRONMENT 28 (5) : 10.1080/23744731.2022.2067466. ScholarBank@NUS Repository. https://doi.org/10.1080/23744731.2022.2067466
Abstract: Research is needed to explore the limitations and potential for improvement of machine learning for building energy prediction. With this aim, the ASHRAE Great Energy Predictor III (GEPIII) Kaggle competition was launched in 2019. This effort was the largest building energy meter machine learning competition of its kind, with 4370 participants who submitted 39,403 predictions. The test dataset included two years of hourly whole building readings from 2380 meters in 1448 buildings at 16 locations. This paper analyzes the various sources and types of residual model error from an aggregation of the competition’s top 50 solutions. This analysis reveals the limitations for machine learning using the standard model inputs of historical meter, weather, and basic building metadata. The errors are classified according to timeframe, behavior, magnitude, and incidence in single buildings or across a campus. The results show machine learning models have errors within a range of acceptability (RMSLEscaled = < 0.1) on 79.1% of the test data. Lower magnitude (in-range) model errors (0.1 < RMSLEscaled = < 0.3) occur in 16.1% of the test data. These errors could be remedied using innovative training data from onsite and Web-based sources. Higher magnitude (out-of-range) errors (RMSLEscaled > 0.3) occur in 4.8% of the test data and are unlikely to be accurately predicted.
Source Title: SCIENCE AND TECHNOLOGY FOR THE BUILT ENVIRONMENT
URI: https://scholarbank.nus.edu.sg/handle/10635/229407
ISSN: 23744731
2374474X
DOI: 10.1080/23744731.2022.2067466
Appears in Collections:Staff Publications
Elements

Show full item record
Files in This Item:
File Description SizeFormatAccess SettingsVersion 
Limitations of machine learning for building energy prediction ASHRAE Great Energy Predictor III Kaggle competition error analysis (1).pdfPublished version4.2 MBAdobe PDF

CLOSED

None

Page view(s)

31
checked on Sep 29, 2022

Download(s)

2
checked on Sep 29, 2022

Google ScholarTM

Check

Altmetric


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.