Comparative Analysis of Random Forest, SVM-RBF, and XGBoost  with SHAP Interpretation for Google Earth Engine-Based Landslide Susceptibility Classification in Wonosobo Regency

Mizana Amaliya; Erna Dwi Astuti; Rina Mahmudati

doi:10.24252/instek.v11i2.68406

Authors

Mizana Amaliya Universitas Sains Al Qur'an
Erna Dwi Astuti Universitas Sains Al Qur'an
Rina Mahmudati Universitas Sains Al Qur'an

DOI:

https://doi.org/10.24252/instek.v11i2.68406

Keywords:

geospatial, landslide susceptibility, machine learning, Wonosobo, XGBoost

Abstract

Landslides constitute the predominant disaster type in Wonosobo Regency, Central Java, with 190 recorded events (62% of all disaster incidents) in 2024. Prior studies have reported inconsistent performance among machine learning algorithms due to variations in geographic context, dataset characteristics, feature selection, and evaluation frameworks. A review of the literature reveals a lack of comparative evaluations of Random Forest (RF), Support Vector Machine with Radial Basis Function kernel (SVM-RBF), and XGBoost that integrate Google Earth Engine (GEE), SHAP interpretability analysis, and three-class susceptibility classification specifically for Wonosobo Regency. This study aims to evaluate and compare the performance of these three algorithms in mapping landslide susceptibility using geospatial data. A total of 642 sample points (321 landslide and 321 non-landslide locations) were analyzed using eight conditioning factors extracted via GEE, seven of which were retained following Variance Inflation Factor (VIF) testing. Model performance was evaluated using accuracy, Cohen’s Kappa, Matthews Correlation Coefficient (MCC), and AUC-ROC with 10-fold stratified cross-validation. XGBoost achieved the highest performance on the study dataset, with an accuracy of 93.02%, an AUC-ROC of 0.974, and a cross-validation accuracy of 91.0% ± 3.1%. SHAP analysis indicated that rainfall made the largest contribution to the model predictions, highlighting its importance in the model’s decision-making process. Classification of prediction probabilities at the sample points using the Jenks Natural Breaks method yielded a susceptibility distribution of 49.2% Low, 6.4% Moderate, and 44.4% High.

Downloads

Download data is not yet available.

References

[1] S. Fidan, H. Tanyas, A. Akbas, L. Lombardo, D. N. Petley, and T. Görüm, “Understanding fatal landslides at global scales: a summary of topographic, climatic, and anthropogenic perspectives,” Nat. Hazards, vol. 120, no. 7, pp. 6437–6455, 2024, doi: 10.1007/s11069-024-06487-3.

[2] M. S. Lanto, D. N. Syam, A. Wahyuni, and A. Kusmiran, “Pemetaan Risiko Bencana Tanah Longsor Di Kabupaten Barru Menggunakan Metode Pembobotan Parameter Dan Inarisk Bnpb,” J. Geosaintek, vol. 8, no. 3, p. 250, 2022, doi: 10.12962/j25023659.v8i3.14758.

[3] M. R. B. E. Prasetyo, “BNPB Butuh Kesepahaman Perkuat Mitigasi Bencana Tanah Longsor,” 2025. Accessed: May 09, 2026. [Online]. Available: https://www.antaranews.com/berita/4874533/bnpb-butuh-kesepahaman-perkuat-mitigasi-bencana-tanah-longsor

[4] A. Sugiharto, “Integrasi Iman dan Ekologi : Analisis Ekoteologi Kristen Terhadap Banjir dan Tanah Longsor 2025,” Vol. 10, No. 1, 2026.

[5] H. M. Darojati, “Analisis Kesesuaian Sebaran Permukiman Dalam Upaya Mitigasi Bencana Tanah Longsor (Studi di Desabinangun, Kecamatan Watumalang, Kabupaten Wonosobo),” Sekolah Tinggi Pertanahan Nasional Yogyakarta, 2024.

[6] M. Fakhri and S. S. Jumadi, “Identifikasi Lokasi Potensial Untuk Relokasi Permukiman Rawan Bencana Longsor di Kabupaten Wonosobo,” 2023, [Online]. Available: https://eprints.ums.ac.id/id/eprint/108037

[7] T. A. Ndofah and P. B. Santosa, “Evaluasi Penggunaan Lahan Mengacu pada Indeks Potensi Lahan dan Kesesuaiannya Terhadap Rencana Tata Ruang Wilayah di Kabupaten Wonosobo,” JGISE J. Geospatial Inf. Sci. Eng., vol. 6, no. 2, p. 87, 2023, doi: 10.22146/jgise.91079.

[8] “Infografis Kejadian Bencana Tahun 2024,” 2025.

[9] P. K. Badapalli et al., “Landslide Susceptibility Level Mapping in Kozhikode, Kerala, Using Machine Learning-Based Random Forest, Remote Sensing, and GIS Techniques,” Land, vol. 14, no. 7, pp. 1–25, 2025, doi: 10.3390/land14071453.

[10] A. Mani, M. Kumari, and R. Badola, “Landslide hazard zonation (LHZ) mapping of Doon Valley using multi-criteria analysis method based on remote sensing and GIS techniques,” Discov. Geosci., vol. 2, no. 1, 2024, doi: 10.1007/s44288-024-00044-y.

[11] M. A. Haq et al., “Analysis of environmental factors using AI and ML methods,” Sci. Rep., vol. 12, no. 1, pp. 1–16, 2022, doi: 10.1038/s41598-022-16665-7.

[12] X. H. Le, C. Choi, S. Eu, M. Yeon, and G. Lee, “Quantitative evaluation of uncertainty and interpretability in machine learning-based landslide susceptibility mapping through feature selection and explainable AI,” Front. Environ. Sci., vol. 12, no. July, pp. 1–15, 2024, doi: 10.3389/fenvs.2024.1424988.

[13] K. Halder et al., “Improving landslide susceptibility prediction through ensemble recursive feature elimination and meta-learning framework,” Sci. Rep., vol. 15, no. 1, pp. 1–20, 2025, doi: 10.1038/s41598-025-87587-3.

[14] J. Yao, X. Yao, Z. Zhao, and X. Liu, “Performance comparison of landslide susceptibility mapping under multiple machine-learning based models considering InSAR deformation: a case study of the upper Jinsha River,” Geomatics, Nat. Hazards Risk, vol. 14, no. 1, p., 2023, doi: 10.1080/19475705.2023.2212833.

[15] Z. Zhao, Z. yuan Liu, and C. Xu, “Slope Unit-Based Landslide Susceptibility Mapping Using Certainty Factor, Support Vector Machine, Random Forest, CF-SVM and CF-RF Models,” Front. Earth Sci., vol. 9, no. March, pp. 1–16, 2021, doi: 10.3389/feart.2021.589630.

[16] K. L. Du, R. Zhang, B. Jiang, J. Zeng, and J. Lu, “Foundations and Innovations in Data Fusion and Ensemble Learning for Effective Consensus,” Mathematics, vol. 13, no. 4, pp. 1–49, 2025, doi: 10.3390/math13040587.

[17] K. L. Du, B. Jiang, J. Lu, J. Hua, and M. N. S. Swamy, “Exploring Kernel Machines and Support Vector Machines: Principles, Techniques, and Future Directions,” Mathematics, vol. 12, no. 24, pp. 1–58, 2024, doi: 10.3390/math12243935.

[18] Z. Rais, S. Muhammad Fahmuddin, Saida, and A. T. Utomo, “Implementation of Machine Learning Algorithm with Extreme Gradient Boosting (XGBoost) Method in Hypertension Level Classification,” J. Appl. Sci. Eng. Technol. Educ., vol. 7, no. 1, pp. 126–136, 2025, doi: 10.35877/454RI.asci4191.

[19] N. Widyaningrum, A. Fudholi, and E. Prawita Setyowati, “The Effect of Altitude Against Total Phenolic and Epigallocatechin Gallate (EGCG) Content in Green Tea Leaves,” Proc. IConSSE FSM SWCU, pp. 34–40, 2015.

[20] A. Addis, “GIS-Based Landslide Susceptibility Mapping Using Frequency Ratio and Shannon Entropy Models in Dejen District, Northwestern Ethiopia,” J. Eng. (United Kingdom), vol. 2023, pp. 115–128, 2023, doi: 10.1155/2023/1062388.

[21] P. Pérez-Cutillas, A. Pérez-Navarro, C. Conesa-García, D. A. Zema, and J. P. Amado-Álvarez, “What is going on within google earth engine? A systematic review and meta-analysis,” Remote Sens. Appl. Soc. Environ., vol. 29, no. September 2022, p. 100907, 2023, doi: 10.1016/j.rsase.2022.100907.

[22] A. Velastegui-Montoya, N. Montalván-Burbano, P. Carrión-Mero, H. Rivera-Torres, L. Sadeck, and M. Adami, “Google Earth Engine: A Global Analysis and Future Trends,” Remote Sens., vol. 15, no. 14, 2023, doi: 10.3390/rs15143675.

[23] B. Alshehri, Z. Zhang, and X. Liu, “A Review of Google Earth Engine for Land Use and Land Cover Change Analysis: Trends, Applications, and Challenges,” ISPRS Int. J. Geo-Information, vol. 14, no. 11, pp. 1–26, 2025, doi: 10.3390/ijgi14110416.

[24] M. Vizzari, “Object-Based Land Cover Classification in Google Earth Engine,” Remote Sens., vol. 14, pp. 1–19, 2022.

[25] K. Maharana, S. Mondal, and B. Nemade, “A review: Data pre-processing and data augmentation techniques,” Glob. Transitions Proc., vol. 3, no. 1, pp. 91–99, 2022, doi: 10.1016/j.gltp.2022.04.020.

[26] Q. Zhao and C. Poudyal, “Credibility theory based on winsorizing,” Eur. Actuar. J., vol. 14, no. 3, pp. 777–807, 2024, doi: 10.1007/s13385-024-00391-7.

[27] Y. W. Rabby, Y. Li, and H. Hilafu, “An objective absence data sampling method for landslide susceptibility mapping,” Sci. Rep., vol. 13, no. 1, pp. 1–15, 2023, doi: 10.1038/s41598-023-28991-5.

[28] J. Kotzé, J. Le Roux, and J. van Tol, “Investigating Soil Properties at Landslide Locations in the Eastern Cape Province, South Africa,” GeoHazards, vol. 6, no. 4, pp. 1–16, 2025, doi: 10.3390/geohazards6040068.

[29] R. Gui et al., “Digital Elevation Model-Driven River Channel Boundary Monitoring Using the Natural Breaks (Jenks) Method,” Remote Sens., vol. 17, no. 6, pp. 1–24, 2025, doi: 10.3390/rs17061092.

[30] M. Azad, T. H. Nehal, and M. Moshkov, A novel ensemble learning method using majority based voting of multiple selective decision trees, vol. 107, no. 1. Springer Vienna, 2025. doi: 10.1007/s00607-024-01394-8.

[31] A. Roy and S. Chakraborty, “Support vector machine in structural reliability analysis: A review,” Reliab. Eng. Syst. Saf., vol. 233, no. April, 2023, doi: 10.1016/j.ress.2023.109126.

[32] M. Wiens, A. Verone-Boyle, N. Henscheid, J. T. Podichetty, and J. Burton, “A Tutorial and Use Case Example of the eXtreme Gradient Boosting (XGBoost) Artificial Intelligence Algorithm for Drug Development Applications,” Clin. Transl. Sci., vol. 18, no. 3, 2025, doi: 10.1111/cts.70172.

[33] M. Castejón-Limas, L. Fernández-Robles, H. Alaiz-Moretón, J. Cifuentes-Rodriguez, and C. Fernández-Llamas, “A Framework for the Optimization of Complex Cyber-Physical Systems via Directed Acyclic Graph,” Sensors, vol. 22, no. 4, 2022, doi: 10.3390/s22041490.

[34] K. M. Sujon, R. Hassan, K. Choi, and M. A. Samad, “Accuracy, precision, recall, f1-score, or MCC? empirical evidence from advanced statistics, ML, and XAI for evaluating business predictive models,” J. Big Data, vol. 12, no. 1, 2025, doi: 10.1186/s40537-025-01313-4.

[35] D. Chicco and G. Jurman, “The Matthews correlation coefficient (MCC) should replace the ROC AUC as the standard metric for assessing binary classification,” BioData Min., vol. 16, no. 1, pp. 1–23, 2023, doi: 10.1186/s13040-023-00322-4.

[36] D. Khan, W. Akram, and S. Ullah, “Enhancing landslide susceptibility predictions with XGBoost and SHAP: a data-driven explainable AI method,” Geocarto Int., vol. 40, no. 1, p., 2025, doi: 10.1080/10106049.2025.2514725.