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

2019 ◽  
Vol 8 (9) ◽  
pp. 385 ◽  
Author(s):  
Emmanuel Papadakis ◽  
Song Gao ◽  
George Baryannis
Keyword(s):  
Los Angeles ◽  
Topic Modeling ◽  
Design Patterns ◽  
Latent Dirichlet Allocation ◽  
Expert Knowledge ◽  
Urban Setting ◽  
Data Driven ◽  
Knowledge Based ◽  
Functional Regions ◽  
Data Driven Approach

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.

scholarly journals Fusing Knowledge-Based and Data-Driven Techniques for the Identification of Urban Functional Regions

2019 ◽  
Author(s):  
Emmanuel Papadakis ◽  
Song Gao ◽  
George Baryannis
Keyword(s):  
Los Angeles ◽  
Topic Modeling ◽  
Latent Dirichlet Allocation ◽  
Expert Knowledge ◽  
Urban Setting ◽  
Data Driven ◽  
Knowledge Based ◽  
Functional Regions ◽  
Planning And Design ◽  
Urban Planning And Design

The problem of identifying functional regions in an urban setting has been approached in literature using two general methodologies: top-down, encoding expert knowledge on urban planning and design (e.g. into patterns) and using that knowledge for identification, and bottom-up, relying on crowdsourcing and Volunteered Geographic Information (VGI) to train learning models, using techniques such as Latent Dirichlet Allocation (LDA) topic modeling. Both approaches have their advantages but also face important 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 data and knowledge in three different ways: functional regions identified from individual approaches are evaluated against each other, knowledge from patterns is used to adjust learning model results and topic models are used to adjust pattern-based results. The proposed methodologies are demonstrated through the use case of identifying shopping-related functional regions in the Los Angeles metropolitan area. Results show that the combination of results from knowledge-based and data-driven techniques can help uncover discrepancies between the two different approaches and smoothen inaccuracies caused by the limitations of each approach.

scholarly journals Who Are the Intended Users of CSR Reports? Insights from a Data-Driven Approach

2021 ◽  
Vol 13 (3) ◽  
pp. 1070
Author(s):  
Charlie Lindgren ◽  
Asif M. Huq ◽  
Kenneth Carling
Keyword(s):  
Topic Modeling ◽  
Latent Dirichlet Allocation ◽  
Accounting Standards ◽  
Data Driven ◽  
Learning Approach ◽  
Holistic View ◽  
Machine Learning Approach ◽  
Data Driven Approach ◽  
Corporate Social ◽  
Bayesian Machine Learning

There is extant research on theorization, conceptualization, determinants, and consequences of corporate social responsibility (CSR). However, what firms include in their CSR or sustainability reports are much less covered and are predominantly covered in case studies of individual firms. In this paper, we instead take a holistic view and simultaneously explore what firms around the globe currently disclose in these reports, more specifically we investigate if firms are shareholder or stakeholder focused. In this investigation, we check the alignment of the reports to the materiality framework of Sustainability Accounting Standards Board (SASB) which was developed having shareholders as the intended user. To estimate what firms disclose in CSR reports we used the unsupervised Bayesian machine learning approach latent Dirichlet allocation (LDA) developed by Blei et al. We conclude that firms target shareholders as the intended users of these reports, even in environments where stakeholder approach of management is argued to be more dominant. Methodologically, we contribute by demonstrating that topic modeling can enhance the objectivity in reviewing CSR-reports.

scholarly journals Quantitative evaluation of ontology design patterns for combining pathology and anatomy ontologies

2018 ◽  
Author(s):  
Sarah M. Alghamdi ◽  
Beth A. Sundberg ◽  
John P. Sundberg ◽  
Paul N. Schofield ◽  
Robert Hoehndorf
Keyword(s):  
Semantic Similarity ◽  
Quantitative Evaluation ◽  
Design Patterns ◽  
Data Driven ◽  
Anatomic Pathology ◽  
Ontology Design ◽  
Data Driven Approach ◽  
Rich Information ◽  
The Subject ◽  
Multiple Ontologies

ABSTRACTData are increasingly annotated with multiple ontologies to capture rich information about the features of the subject under investigation. Analysis may be performed over each ontology separately, but, recently, there has been a move to combine multiple ontologies to provide more powerful analytical possibilities. However, it is often not clear how to combine ontologies or how to assess or evaluate the potential design patterns available. Here we use a large and well-characterized dataset of anatomic pathology descriptions from a major study of aging mice. We show how different design patterns based on the MPATH and MA ontologies provide orthogonal axes of analysis, and perform differently in over-representation and semantic similarity applications. We discuss how such a data-driven approach might be used generally to generate and evaluate ontology design patterns.

Sign in / Sign up

Export Citation Format

Share Document