Explainable artificial intelligence empowered landslide susceptibility mapping using Extreme Gradient Boosting (XGBoost)
Keywords:
Landslide susceptibility, XAI, SHAP, XGBoost, SapancaAbstract
Up to now, a wide variety of non-linear machine learning models with black-box nature have been intensively utilized to spatially predict landslide susceptibility in a given geographical context. However, the results obtained from these models can be difficult to interpret, making it challenging to identify the reasons for false positives and take corrective action. To address this problem, this study makes use of the XGBoost algorithm to predict landslide susceptibility in a lake basin and its surrounding areas. Additionally, the Shapley additive explanation (SHAP) approach as an explainable artificial intelligence (XAI) tool was used to increase the interpretability of the model’s predictions. The accuracy of the XGBoost model was evaluated and found to have an OA of 92.44% and an AUC score of 98.73%. The SHAP analysis showed that slope was the most influential factor in predicting landslide susceptibility. Additionally, the dependence plot highlighted that the impact of slope angle on the model's output was consistent within the range of 8° to 21°. The findings of this study demonstrate the potential benefits of incorporating XAI techniques into the modeling process to increase transparency.
References
Kavzoglu, T. (2009). Increasing the accuracy of neural network classification using refined training data. Environmental Modelling & Software, 24(7), 850–858. doi: https://doi.org/10.1016/j.envsoft.2008.11.012
Kavzoglu, T., & Teke, A. (2022). Advanced hyperparameter optimization for improved spatial prediction of shallow landslides using extreme gradient boosting (XGBoost). Bulletin of Engineering Geology and the Environment, 81(5), 201. https://doi.org/10.1007/s10064-022-02708-w
Chen, T., & Guestrin, C. (2016). XGBoost: A Scalable Tree Boosting System. In Proceedings of 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, CA, USA, 13–17 August, 785–794.
Lundberg, S. M., & Lee, S. I. (2017). A unified approach to interpreting model predictions. Advances in Neural Information Processing Systems, Long Beach, CA, USA, 4–9 December, 4766–4775.
Kavzoglu, T., & Teke, A. (2022). Predictive Performances of Ensemble Machine Learning Algorithms in Landslide Susceptibility Mapping Using Random Forest, Extreme Gradient Boosting (XGBoost) and Natural Gradient Boosting (NGBoost). Arabian Journal for Science and Engineering, 47, 7367–7385. https://doi.org/ 10.1007/s13369-022-06560-8
