Interpretable Landslide Susceptibility Mapping in Pacitan, Indonesia Using Bayesian-Optimized XGBoost and SHAP Analysis
DOI:
https://doi.org/10.59261/jequi.v8i5.459Keywords:
Landslide Susceptibility Mapping, XGBoost, Bayesian Optimization, SHAP, QGIS, PacitanAbstract
Background: Landslide susceptibility mapping is essential in areas with complex topographic and geological conditions, such as Pacitan Regency, Indonesia, to identify landslide-prone areas and assess the factors influencing landslide susceptibility.
Objective: This study developed a landslide susceptibility mapping model using Extreme Gradient Boosting (XGBoost), optimized through Bayesian optimization and interpreted using the SHapley Additive exPlanations (SHAP) method.
Methods: The dataset consisted of 1,486 samples, comprising 743 landslide occurrence points and 743 non-landslide points. Eleven landslide conditioning factors were used as predictor variables: elevation, aspect, slope, geology, land use, profile curvature, plan curvature, distance to roads, distance to rivers, Stream Power Index (SPI), and Topographic Wetness Index (TWI). Model hyperparameter optimization was performed using Bayesian optimization with a Tree-structured Parzen Estimator (TPE).
Results: The optimized XGBoost model achieved an accuracy of 69.80%, a precision of 70.34%, a recall of 68.46%, an F1 score of 69.39%, and a receiver operating characteristic area under the curve (ROC-AUC) of 0.7564 on the independent test set. The resulting susceptibility map classified the study area into five categories: very low, low, moderate, high, and very high susceptibility. SHAP analysis identified slope, aspect, plan curvature, geology, and distance to roads as the dominant factors influencing model predictions.
Conclusion: Integrating Bayesian optimization, XGBoost, and SHAP provides an interpretable approach to landslide susceptibility mapping in Pacitan Regency. The resulting susceptibility map and feature-level explanations can support preliminary disaster risk assessment and spatial planning. However, residual classification errors and limitations associated with the use of static environmental predictors should be considered when interpreting the results.
Downloads
References
Ado, M., Amitab, K., Maji, A. K., Jasińska, E., Gono, R., Leonowicz, Z., & Jasiński, M. (2022). Landslide susceptibility mapping using machine learning: A literature survey. Remote Sensing, 14(13), 3029. https://doi.org/10.3390/rs14133029
Akiba, T., Sano, S., Yanase, T., Ohta, T., & Koyama, M. (2019). Optuna: A next-generation hyperparameter optimization framework. Dalam Proceedings of the 25th ACM SIGKDD International Conference on Knowledge Discovery & Data Mining (2623–2631). https://doi.org/10.1145/3292500.3330701
Bergstra, J., Bardenet, R., Bengio, Y., & Kégl, B. (2011). Algorithms for hyper-parameter optimization. Dalam Advances in Neural Information Processing Systems (Vol. 24, hlm. 2546–2554).
Bergstra, J., & Bengio, Y. (2012). Random search for hyper-parameter optimization. Journal of Machine Learning Research, 13, 281–305.
Bischl, B., Binder, M., Lang, M., Pielok, T., Richter, J., Coors, S., Thomas, J., Ullmann, T., Becker, M., Boulesteix, A.-L., Deng, D., & Lindauer, M. (2021). Hyperparameter optimization: Foundations, algorithms, best practices, and open challenges. WIREs Data Mining and Knowledge Discovery, 11(2), e1484. https://doi.org/10.1002/widm.1484
Chen, T., & Guestrin, C. (2016). XGBoost: A scalable tree boosting system. Dalam Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (785–794). https://doi.org/10.1145/2939672.2939785
Darminto, M. R., Widodo, A., Alfatinah, A., & Chu, H. J. (2021). High-resolution landslide susceptibility map generation using machine learning (case study in Pacitan, Indonesia). International Journal on Advanced Science, Engineering and Information Technology, 11(1), 369–379. https://doi.org/10.18517/ijaseit.11.1.11679
Dasgupta, S., & Sen, J. (2024). A comparative study of hyperparameter tuning methods. Data Science in Theory and Practice, 258.
Dimyati, M., Ash Shidiq, I. P., & Ichsandya, D. B. (2022). GIS-based binary logistic regression for landslide susceptibility mapping in the central part of Pacitan Regency, East Java, Indonesia. International Journal on Advanced Science, Engineering and Information Technology, 12(4), 1462–1468. https://doi.org/10.18517/ijaseit.12.4.16221
Grandini, M., Bagli, E., & Visani, G. (2020). Metrics for multi-class classification: an overview. arXiv preprint arXiv:2008.05756.
Guidotti, R., Monreale, A., Ruggieri, S., Turini, F., Giannotti, F., & Pedreschi, D. (2018). A survey of methods for explaining black box models. ACM Computing Surveys, 51(5), Article 93. https://doi.org/10.1145/3236009
Hussain, M. A., Chen, Z., Shoaib, M., et al. (2022). Landslide susceptibility mapping using machine learning algorithm validated by persistent scatterer InSAR technique. Sensors, 22(9), 3119. https://doi.org/10.3390/s22093119
Khan Inan, M. S., & Rahman, I. (2022). Explainable AI integrated feature selection for landslide susceptibility mapping using TreeSHAP. arXiv. https://doi.org/10.1007/s42979-023-01960-5
Kotsiantis, S. B., Kanellopoulos, D., & Pintelas, P. E. (2006). Data preprocessing for supervised learning. International Journal of Computer Science, 1(2), 111–117.
Krawczyk, B. (2016). Learning from imbalanced data: Open challenges and future directions. Progress in Artificial Intelligence, 5(4), 221–232. https://doi.org/10.1007/s13748-016-0094-0
Lundberg, S. M., Erion, G., Chen, H., DeGrave, A., Prutkin, J. M., Nair, B., Katz, R., Himmelfarb, J., Bansal, N., & Lee, S.-I. (2020). From local explanations to global understanding with explainable AI for trees. Nature Machine Intelligence, 2(1), 56–67. https://doi.org/10.1038/s42256-019-0138-9
Lundberg, S. M., & Lee, S.-I. (2017). A unified approach to interpreting model predictions. Dalam Advances in Neural Information Processing Systems (Vol. 30).
Molnar, C. (2020). Interpretable machine learning. Lulu. com.
Mulabbi, A., Danoedoro, P., & Samodra, G. (2025). Integrating interpretable artificial neural networks, geomorphic plausibility and transfer learning for landslide susceptibility mapping: A case study of Pacitan, East Java. Discover Sustainability, 6(1). https://doi.org/10.1007/s43621-025-01959-3
Ponce-Bobadilla, A. V., Schmitt, V., Maier, C. S., Mensing, S., & Stodtmann, S. (2024). Practical guide to SHAP analysis: Explaining supervised machine learning model predictions in drug development. Clinical and Translational Science. https://doi.org/10.1111/cts.70056
Putra, A. N., Nita, I., Jauhary, M. R. A., Nurhutami, S. R., & Ismail, M. H. (2021). Landslide risk analysis on agriculture area in Pacitan Regency in East Java Indonesia using geospatial techniques. Environment and Natural Resources Journal, 19(3). https://doi.org/10.32526/ennrj/19/2020167
Samodra, G., Malawani, M. N., Suhendro, I., & Mardiatno, D. (2024). Spatial datasets for benchmarking machine learning-based landslide susceptibility models. Data in Brief, 57, 111155. https://doi.org/10.17632/vrtx3w6mjd.1
Snoek, J., Larochelle, H., & Adams, R. (2012). Practical bayesian optimization of machine learning algorithms. Advances in neural information processing systems, 25.
Sun, D., Chen, D., Zhang, J., Mi, C., Gu, Q., & Wen, H. (2023). Landslide susceptibility mapping based on interpretable machine learning from the perspective of geomorphological differentiation. Land, 12(5), Article 1018. https://doi.org/10.3390/land12051018
Watanabe, S. (2023). Tree-structured Parzen estimator: Understanding its algorithm components and their roles for better empirical performance. arXiv. https://arxiv.org/abs/2304.11127
Downloads
Published
Issue
Section
License
Copyright (c) 2026 Aghasi Yanda Wafa Azizah, Khalila Salma Dzakia Agram, Shellomitha Sulvana Dewi, Saira Nashwa Ramadhan, Yuliani Puji Astuti
This work is licensed under a Creative Commons Attribution-ShareAlike 4.0 International License.
Authors who publish with this journal agree to the following terms:
- Authors retain copyright and grant the journal right of first publication with the work simultaneously licensed under a Creative Commons Attribution-ShareAlike 4.0 International (CC-BY-SA). that allows others to share the work with an acknowledgement of the work's authorship and initial publication in this journal.
- Authors are able to enter into separate, additional contractual arrangements for the non-exclusive distribution of the journal's published version of the work (e.g., post it to an institutional repository or publish it in a book), with an acknowledgement of its initial publication in this journal.
- Authors are permitted and encouraged to post their work online (e.g., in institutional repositories or on their website) prior to and during the submission process, as it can lead to productive exchanges, as well as earlier and greater citation of published work.
