Keywords
Land use change; Cellular automata; support vector machines; random forest
Start Date
5-7-2022 12:00 PM
End Date
8-7-2022 9:59 AM
Abstract
Land-use change (LUC) models are important tools that provide insight into urbanization dynamics and possible future patterns. Many domains employ LUC models to project the possible impact of changes in land use on, e.g., ecosystem services, weather-related extremes such as flooding and heat. The calibration process is the core theme of LUC models. This study compares the performance of two common machine-learning classifiers, random forest (RF), and support vector machines (SVM), to calibrate Cellular automata LUC models (CA). It focuses on the sensitivity analysis of the sample size and the number of input variables for each classifier. We applied the models to the Wallonia region (Belgium) as a case study to demonstrate the performance of each classifier. The evaluation process of the performance of our model is based on the comparison of the simulated built-up maps of 2010 with that produced by the model with the real map of 2010 using the fuzzy similarity rate. Our results highlight that RF produces a land-use pattern that simulates the observed pattern more precisely than SVM, especially with a low sample size, which is important for study areas with low levels of land-use change. Although zoning information notably enhances the accuracy of SVM-based probability maps, zoning marginally influences the RF-derived probability maps. In the case of the SVM, the CA model did not significantly improve due to the increased sample size. The RF-driven CA had the best performance with a high sample, while zoning information was excluded.
The value of AI in land-use change models
Land-use change (LUC) models are important tools that provide insight into urbanization dynamics and possible future patterns. Many domains employ LUC models to project the possible impact of changes in land use on, e.g., ecosystem services, weather-related extremes such as flooding and heat. The calibration process is the core theme of LUC models. This study compares the performance of two common machine-learning classifiers, random forest (RF), and support vector machines (SVM), to calibrate Cellular automata LUC models (CA). It focuses on the sensitivity analysis of the sample size and the number of input variables for each classifier. We applied the models to the Wallonia region (Belgium) as a case study to demonstrate the performance of each classifier. The evaluation process of the performance of our model is based on the comparison of the simulated built-up maps of 2010 with that produced by the model with the real map of 2010 using the fuzzy similarity rate. Our results highlight that RF produces a land-use pattern that simulates the observed pattern more precisely than SVM, especially with a low sample size, which is important for study areas with low levels of land-use change. Although zoning information notably enhances the accuracy of SVM-based probability maps, zoning marginally influences the RF-derived probability maps. In the case of the SVM, the CA model did not significantly improve due to the increased sample size. The RF-driven CA had the best performance with a high sample, while zoning information was excluded.
Stream and Session
false