Comparative Analysis of Machine Learning Algorithms for Building Archetypes Development in Urban Building Energy Modeling

Files in This Item:
File Description SizeFormat 
2018_Ali_Clustering_SimBuild.pdf570.51 kBAdobe PDFDownload
Title: Comparative Analysis of Machine Learning Algorithms for Building Archetypes Development in Urban Building Energy Modeling
Authors: Ali, UsmanShamsi, Mohammad HarisAlshehri, FawazMangina, EleniO'Donnell, James
Permanent link:
Date: 28-Sep-2018
Online since: 2019-08-21T06:59:03Z
Abstract: The most common approach for urban building energy modeling (UBEM) involves segmenting a building stock into archetypes. Development Building archetypes for urban scale is a complex task and requires a lot of extensive data. The archetype development methodology proposed in this paper uses unsupervised machine learning approaches to identify similar clusters of buildings based on building specific features. The archetype development process considers four crucial processes of machine learning: data preprocessing, feature selection, clustering algorithm adaptation and results validation. The four different clustering algorithms investigated in this study are KMean, Hierarchical, Density-based, K-Medoids. All the algorithms are applied on Irish Energy Performance Certificate (EPC) that consist of 203 features. The obtained results are then used to compare and analyze the chosen algorithms with respect to performance, quality and cluster instances. The K-mean algorithm preforms the best in terms of cluster formation.
Funding Details: Science Foundation Ireland
Type of material: Conference Publication
Start page: 60
End page: 67
Copyright (published version): 2018 ASHRAE
Keywords: Urban building energy modeling (UBEM)Archetype development methodologyEnergy Performance Certificate (EPC)Ireland
Other versions:
Language: en
Status of Item: Peer reviewed
Is part of: 2018 Building Performance Analysis Conference and SimBuild
Conference Details: 2018 Building Performance Modeling Conference and SimBuild co-organized by ASHRAE and IBPSA-USA Chicago, IL, 26-28 September 2018
Appears in Collections:Mechanical & Materials Engineering Research Collection
Computer Science Research Collection

Show full item record

Google ScholarTM


This item is available under the Attribution-NonCommercial-NoDerivs 3.0 Ireland. No item may be reproduced for commercial purposes. For other possible restrictions on use please refer to the publisher's URL where this is made available, or to notes contained in the item itself. Other terms may apply.