Combining Design Patterns and Topic Modeling to Discover Regions That Support Particular Functionality

Emmanuel Papadakis, Song Gao, George Baryannis

Research output: Contribution to journalArticle

Abstract

The problem of discovering regions that support particular functionalities in an urban setting has been approached in literature using two general methodologies: top-down, encoding expert knowledge on urban planning and design and discovering regions that conform to that knowledge; and bottom-up, using data to train machine learning models, which can discover similar regions. Both methodologies face limitations, with knowledge-based approaches being criticized for scalability and transferability issues and data-driven approaches for lacking interpretability and depending heavily on data quality. To mitigate these disadvantages, we propose a novel framework that fuses a knowledge-based approach using design patterns and a data-driven approach using latent Dirichlet allocation (LDA) topic modeling in three different ways: Functional regions discovered using either approach are evaluated against each other to identify cases of significant agreement or disagreement; knowledge from patterns is used to adjust topic probabilities in the learning model; and topic probabilities are used to adjust pattern-based results. The proposed methodologies are demonstrated through the use case of identifying shopping-related regions in the Los Angeles metropolitan area. Results show that the combination of pattern-based discovery and topic modeling extraction helps uncover discrepancies between the two approaches and smooth inaccuracies caused by the limitations of each approach.
LanguageEnglish
Article number385
Pages1-22
Number of pages22
JournalISPRS International Journal of Geo-Information
Volume8
Issue number9
DOIs
Publication statusPublished - 3 Sep 2019

Fingerprint

functionality
Urban planning
Electric fuses
modeling
Learning systems
Scalability
methodology
expert knowledge
data quality
urban planning
urban design
knowledge
learning
agglomeration area
metropolitan area
train

Cite this

@article{a802874e950f4b7e9ecaad4f673d7f24,
title = "Combining Design Patterns and Topic Modeling to Discover Regions That Support Particular Functionality",
abstract = "The problem of discovering regions that support particular functionalities in an urban setting has been approached in literature using two general methodologies: top-down, encoding expert knowledge on urban planning and design and discovering regions that conform to that knowledge; and bottom-up, using data to train machine learning models, which can discover similar regions. Both methodologies face limitations, with knowledge-based approaches being criticized for scalability and transferability issues and data-driven approaches for lacking interpretability and depending heavily on data quality. To mitigate these disadvantages, we propose a novel framework that fuses a knowledge-based approach using design patterns and a data-driven approach using latent Dirichlet allocation (LDA) topic modeling in three different ways: Functional regions discovered using either approach are evaluated against each other to identify cases of significant agreement or disagreement; knowledge from patterns is used to adjust topic probabilities in the learning model; and topic probabilities are used to adjust pattern-based results. The proposed methodologies are demonstrated through the use case of identifying shopping-related regions in the Los Angeles metropolitan area. Results show that the combination of pattern-based discovery and topic modeling extraction helps uncover discrepancies between the two approaches and smooth inaccuracies caused by the limitations of each approach.",
keywords = "Functional region, Place, Patterns, Topic modeling, urban planning, Volunteered Geographic Information (VGI)",
author = "Emmanuel Papadakis and Song Gao and George Baryannis",
year = "2019",
month = "9",
day = "3",
doi = "10.3390/ijgi8090385",
language = "English",
volume = "8",
pages = "1--22",
journal = "ISPRS International Journal of Geo-Information",
issn = "2220-9964",
publisher = "MDPI AG",
number = "9",

}

Combining Design Patterns and Topic Modeling to Discover Regions That Support Particular Functionality. / Papadakis, Emmanuel; Gao, Song; Baryannis, George.

In: ISPRS International Journal of Geo-Information, Vol. 8, No. 9, 385, 03.09.2019, p. 1-22.

Research output: Contribution to journalArticle

TY - JOUR

T1 - Combining Design Patterns and Topic Modeling to Discover Regions That Support Particular Functionality

AU - Papadakis, Emmanuel

AU - Gao, Song

AU - Baryannis, George

PY - 2019/9/3

Y1 - 2019/9/3

N2 - The problem of discovering regions that support particular functionalities in an urban setting has been approached in literature using two general methodologies: top-down, encoding expert knowledge on urban planning and design and discovering regions that conform to that knowledge; and bottom-up, using data to train machine learning models, which can discover similar regions. Both methodologies face limitations, with knowledge-based approaches being criticized for scalability and transferability issues and data-driven approaches for lacking interpretability and depending heavily on data quality. To mitigate these disadvantages, we propose a novel framework that fuses a knowledge-based approach using design patterns and a data-driven approach using latent Dirichlet allocation (LDA) topic modeling in three different ways: Functional regions discovered using either approach are evaluated against each other to identify cases of significant agreement or disagreement; knowledge from patterns is used to adjust topic probabilities in the learning model; and topic probabilities are used to adjust pattern-based results. The proposed methodologies are demonstrated through the use case of identifying shopping-related regions in the Los Angeles metropolitan area. Results show that the combination of pattern-based discovery and topic modeling extraction helps uncover discrepancies between the two approaches and smooth inaccuracies caused by the limitations of each approach.

AB - The problem of discovering regions that support particular functionalities in an urban setting has been approached in literature using two general methodologies: top-down, encoding expert knowledge on urban planning and design and discovering regions that conform to that knowledge; and bottom-up, using data to train machine learning models, which can discover similar regions. Both methodologies face limitations, with knowledge-based approaches being criticized for scalability and transferability issues and data-driven approaches for lacking interpretability and depending heavily on data quality. To mitigate these disadvantages, we propose a novel framework that fuses a knowledge-based approach using design patterns and a data-driven approach using latent Dirichlet allocation (LDA) topic modeling in three different ways: Functional regions discovered using either approach are evaluated against each other to identify cases of significant agreement or disagreement; knowledge from patterns is used to adjust topic probabilities in the learning model; and topic probabilities are used to adjust pattern-based results. The proposed methodologies are demonstrated through the use case of identifying shopping-related regions in the Los Angeles metropolitan area. Results show that the combination of pattern-based discovery and topic modeling extraction helps uncover discrepancies between the two approaches and smooth inaccuracies caused by the limitations of each approach.

KW - Functional region

KW - Place

KW - Patterns

KW - Topic modeling

KW - urban planning

KW - Volunteered Geographic Information (VGI)

U2 - 10.3390/ijgi8090385

DO - 10.3390/ijgi8090385

M3 - Article

VL - 8

SP - 1

EP - 22

JO - ISPRS International Journal of Geo-Information

T2 - ISPRS International Journal of Geo-Information

JF - ISPRS International Journal of Geo-Information

SN - 2220-9964

IS - 9

M1 - 385

ER -