Abstract
The use of machine learning has grown in popularity in various disciplines. Despite the popularity, the apparent ‘black box’ nature of such tools continues to be an area of concern. In this article, we attempt to unravel the complexity of this black box by exploring the use of artificial neural networks (ANNs), coupled with graph theory, to model and interpret the spatial distribution of building damage from extreme wind events at a community level. Structural wind damage is a topic that is mostly well understood for how wind pressure translates to extreme loading on a structure, how debris can affect that loading and how specific social characteristics contribute to the overall population vulnerability. While these themes are widely accepted, they have proven difficult to model in a cohesive manner, which has led primarily to physical damage models considering wind loading only as it relates to structural capacity. We take advantage of this modelling difficulty to reflect on two different ANN models for predicting the spatial distribution of structural damage due to wind loading. Through graph theory analysis, we study the internal patterns of the apparent black box of artificial intelligence of the models and show that social parameters are key to predict structural damage.
Highlights
The creation of predictive models using machine learning has grown in popularity in recent years due to the ability to model complex, nonlinear, relationships
Studies have applied such techniques to enhance modelling, forecasts and potentially predict outcomes 2 or impacts from hazards [1,2,3,4,5,6,7,8]. Each of these studies used artificial neural networks (ANNs), a type of machine learning (ML), in which the ANN is best used for modelling a singular problem, for example, forecasting the overall impact, in terms of economic damage, from a hurricane event [1,9,10]
This approach is typically forgone as the problems usually addressed with ANNs are especially difficult to solve conceptually [12]. What such concerns highlight is a need for model transparency through interpretable ML. This becomes concerning when considering that multiple different ML algorithms could produce acceptable results [12], leading to the question of ‘how do we truly know which model is producing the most conceptually accurate results?’ While there is a large trend in evaluating graph theory models using ANNs, we propose that a way to answer this question is by using graph theory as a means of analysing an ANN’s patterns and neural connections
Summary
The creation of predictive models using machine learning has grown in popularity in recent years due to the ability to model complex, nonlinear, relationships. Studies have applied such techniques to enhance modelling, forecasts and potentially predict outcomes 2 or impacts from hazards [1,2,3,4,5,6,7,8] Each of these studies used artificial neural networks (ANNs), a type of machine learning (ML), in which the ANN is best used for modelling a singular problem, for example, forecasting the overall impact, in terms of economic damage, from a hurricane event [1,9,10]. While the authors demonstrate viability for ANNs to be applied in determining tornadic damage and loss, they do suggest additional research to further refine the approach Their model does not consider structural variables, whereas the field’s current physics-based modelling approach to wind-induced damage (and subsequent loss) is solely structural engineering based. The ANN model discussed uses relevant tornado, societal demographic and structural data to determine a building’s resulting damage state from an extreme wind event
Sign-in/Register to view highlights & summary Sign-in/Register to access full text options 
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a callDisclaimer: All third-party content on this website/platform is and will remain the property of their respective owners and is provided on "as is" basis without any warranties, express or implied. Use of third-party content does not indicate any affiliation, sponsorship with or endorsement by them. Any references to third-party content is to identify the corresponding services and shall be considered fair use under The CopyrightLaw.
