Skip to main content

Feature Extraction and K-means Clustering Approach to Explore Important Features of Urban Identity

Chang, Mei-Chih and Buš, Peter and Schmitt, Gerhard (2018) Feature Extraction and K-means Clustering Approach to Explore Important Features of Urban Identity. In: 2017 16th IEEE International Conference on Machine Learning and Applications (ICMLA). IEEE, pp. 1139-1144. ISBN 978-1-5386-1419-8. E-ISBN 978-1-5386-1418-1. (doi:10.1109/ICMLA.2017.00015)

PDF - Author's Accepted Manuscript
Download (1MB) Preview
[img]
Preview
PDF - Publisher pdf
Restricted to Repository staff only
Contact us about this Publication Download (1MB)
[img]
Official URL
http://dx.doi.org/10.1109/ICMLA.2017.00015

Abstract

Public spaces play an important role in theprocesses of formation, generation and change of urban identity.Under present day conditions, the identities of cities are rapidlydeteriorating and vanishing. Therefore, the importance of urbandesign, which is a means of designing urban spaces and theirphysical and social aspects, is ever growing. This paper proposesa novel methodology by using Principle Component Analysis(PCA) and K-means clustering approach to find importantfeatures of the urban identity from public space. K. Lynch`s work and Space Syntax theory are reconstructed and integratedwith POI (Points of Interest) to quantify the quality of the publicspace. A case study of Zürich city is used to test of theseredefinitions and features of urban identity. The results showthat PCA and K-means clustering approach can identify theurban identity and explore important features. This strategycould help to improve planning and design processes andgeneration of new urban patterns with more appropriate featuresand qualities.

Item Type: Book section
DOI/Identification number: 10.1109/ICMLA.2017.00015
Additional information: Unmapped bibliographic data: LA - en [Field not mapped to EPrints] SE - 1139 [Field not mapped to EPrints]
Uncontrolled keywords: Urban identity, Space syntax, k-means clustering, Feature extraction
Subjects: N Visual Arts > NA Architecture
Divisions: Faculties > Humanities > Architecture
Depositing User: Peter Bus
Date Deposited: 06 Sep 2018 10:19 UTC
Last Modified: 26 Sep 2019 11:27 UTC
Resource URI: https://kar.kent.ac.uk/id/eprint/68960 (The current URI for this page, for reference purposes)
Buš, Peter: https://orcid.org/0000-0002-1730-4559
  • Depositors only (login required):

Downloads

Downloads per month over past year