Predicting hurricane evacuation decisions with interpretable machine learning methods

The present study proposed a novel interpretable machine learning approach to predict household-level evacuation decisions by leveraging easily accessible demographic and resource-related predictors, compared to existing models that mainly rely on psychological factors. An enhanced logistic regression model (that is, an interpretable machine learning approach) was developed for accurate predictions by automatically accounting for nonlinearities and interactions (that is, univariate and bivariate threshold effects). Specifically, nonlinearity and interaction detection were enabled by low-depth decision trees, which offer transparent model structure and robustness.

A survey dataset collected in the aftermath of Hurricanes Katrina and Rita, two of the most intense tropical storms of the last two decades, was employed to test the new methodology. The findings show that, when predicting the households’ evacuation decisions, the enhanced logistic regression model outperformed previous linear models in terms of both model fit and predictive capability. This outcome suggests that our proposed methodology could provide a new tool and framework for emergency management authorities to improve the prediction of evacuation traffic demands in a timely and accurate manner.

Source: International Journal of Disaster Risk Science

Author(s): Yuran SunShih-Kai HuangXilei Zhao

Are you sure you want to delete this "resource"?
This item will be deleted immediately. You cannot undo this action.

Related Resources

Report, Research
06 Apr 2023
In recent years, chatbots have offered humanitarian operations the possibility to automate personalised engagement and support, inform tailored programme design and gather and share information at a large scale. However, adopting a chatbot is never...
Tags: Report, Research, Mobile Technology
Research
31 Oct 2013
The study published by IFRC finds that while social capital has always played an implicit role in the community-based work of the Red Cross/Red Crescent Movement – the volunteerism at its core, as well as the community-based committees and ac...
Tags: Research, Public Awareness and Public Education
Research
06 Jul 2015
Asia and the Pacific is the most disaster-prone region in the world. Geologically, the region is characterized by active tectonic plate movements in the Pacific and Indian Oceans, which have been the source of major earthquakes and tsunamis. In the p...
Tags: Research, Earthquake, Flood, Hurricane / Typhoon / Cyclone, Landslide, Thunderstorm, Tsunami
Scroll to Top