Prediction of Casualty Rate in Emergency Situations Using the XGBoost Method and SHAP Interpretation (Case Study: Upt. Basarnas Medan)
Abstract
This study aims to predict the level of casualties in emergency situations using the XGBoost method and interpret the model results using SHAP (Shapley Additive Explanations). The data used comes from the SAR operation of Basarnas Medan with variables of incident type, response time, number of personnel, and distance of the incident. The XGBoost model is used to classify the level of casualties into three categories, namely low, medium, and high. The test results show that the model achieved an accuracy level of 87,77%, which indicates a fairly good model ability in classifying. Evaluation using a confusion matrix shows that the model has the best performance in categories with a larger amount of data, especially the low category, while in the medium and high categories there are still some classification errors due to similarity of characteristics between classes and data imbalance. The SHAP interpretation analysis shows that the type of incident is the most dominant factor influencing the model prediction with the highest contribution value of 1,55 , followed by the number of personnel (0,71), the distance of the incident (0,68), and the response time (0,62). This finding is also supported by the SHAP Summary Plot visualization which shows that the type of incident has the broadest influence on the prediction results. Overall, this study shows that the XGBoost method is not only capable of producing accurate predictions, but can also be well interpreted through the SHAP approach, thus potentially supporting more effective decision-making in handling emergency conditions to minimize the level of casualties.
References
W. Andrianto, “Jurnal Hukum & Pembangunan Tanggung Jawab Hukum Sumber Daya Manusia Potensi Basarnas Dalam Melakukan Tindakan Medis Terhadap Korban Bencana,” Vol. 51, No. 4, 2021, Doi: 10.21143/Jhp.Vol51.No4.3297.
J. Hukum And I. U. S. Quia, “Https://Doi.Org/10.20885/Iustum.Vol31.Iss1.Art3,” Vol. 31, No. September 2022, Pp. 49–75, 2024.
A. Setiawan, D. Ruslanjari, D. Puspitasari, And N. Pencarian, “Penilaian Risiko Kesehatan Mental Rescuer Dalam Mendukung Operasi Tanggap Darurat Bencana Pendahuluan,” Vol. 2, No. 1, Pp. 1–14, 2024.
A. A. Akbar, H. Dwiningtias, And H. K. Rahmat, “Urgensi Koordinasi Dalam Organisasi Tanggap Darurat Bencana Di Indonesia : Sebuah Tinjauan Pustaka,” Vol. 1, No. 1, Pp. 15–20, 2024.
M. F. A. Burhanuddin, U. Widyaningsih, And T. Rahayu, “Penerapan Latihan Keadaan Darurat Dalam Upaya Persiapan Menangani Keadaan Darurat Di Kapal Politeknik Pelayaran Surabaya , Indonesia,” Vol. 4, Pp. 1362–1378, 2024.
A. Khuznuzzan And D. Widagdo, “Implementasi Peraturan Menteri Perhubungan Nomor 95 Tahun 2021 Mengenai Pelatihan Penanggulangan Keadaan Darurat Di Bandar Udara Internasional Zainuddin Abdul,” No. 1, Pp. 1–9, 2024.
P. M. Bencana, U. Dharmawangsa, And M. Danil, “Manajemen Bencana,” No. November, Pp. 7–14, 2021.
R. Tanggap, D. Gempa, M. Di, And D. Tadui, “Jurnal Pengabdian Kepada Masyarakat,” Vol. 1, 2021, Doi: 10.35329/Sipissangngi.V1i1.1890.
I. Of, D. Policy, D. Response, And I. N. Indonesia, “Integrasi Kebijakan Pertahanan Dan Respon Bencana Di Indonesia ; Pasukan Pe…,” Vol. 7, Pp. 163–175, 2023.
Y. Feng, D. Wang, Y. Yin, Z. Li, And Z. Hu, “An Xgboost ‑ Based Casualty Prediction Method For Terrorist Attacks,” Complex Intell. Syst., Vol. 6, No. 3, Pp. 721–740, 2020, Doi: 10.1007/S40747-020-00173-0.
“머신러닝을 이용한 호우재해의 인명피해 예측모델 개발 Development Of A Machine Learning Model For Predicting Casualties From Heavy Rainfall Disasters,” Vol. 25, No. 5, Pp. 231–240, 2025.
I. M. Akhsan, K. Utami, D. Octavia, And C. Salsabilla, “Eksplorasi Dan Komparasi Model Klasifikasi Machine Learning Untuk Prediksi Tsunami,” Vol. 5, No. 2, Pp. 149–157, 2025.
A. R. Febrian, “Memprediksi Tingkat Kecelakan Jalan Raya Di Salatiga Menggunakan Machine Learning,” Vol. 10, No. 4, Pp. 4274–4283, 2025.
Z. Li, “Computers , Environment And Urban Systems Extracting Spatial Effects From Machine Learning Model Using Local Interpretation Method : An Example Of Shap And Xgboost,” Comput. Environ. Urban Syst., Vol. 96, No. November 2021, P. 101845, 2022, Doi: 10.1016/J.Compenvurbsys.2022.101845.
E. Awareness, I. Promotingsustainable, M. Iqbal, And A. J. Lubis, “Smart Technology Based On Ldr And Smartphone,” Pp. 528–537.
I. M. Faiza And W. Andriani, “Tinjauan Pustaka Sistematis : Penerapan Metode Machine Learning Untuk Deteksi Bencana Banjir,” Vol. 11, No. September, Pp. 59–63, 2022.
K. Metode And X. Dan, “No Title,” Vol. 05, No. 01, Pp. 26–37, 2024.
P. S. Rizky Et Al., “Perbandingan Metode Lightgbm Dan Xgboost Dalam Menangani Data Dengan Kelas Tidak Seimbang Universitas Hamzanwadi Selong,” Vol. 15, No. 2, Pp. 228–236, 2022.
J. Sistem And I. Tgd, “Model Prediksi Penjadwalan Produksi Energi Terbarukan Dengan Algoritma Xgboost Dan Analisis Interpretatif Menggunakan Shap,” Vol. 4, Pp. 794–801, 2025.
A. A. Sutrisno Et Al., “Dengan Model Explainable Berbasis Shap,” Vol. 10, No. 1, Pp. 296–301, 2026.
Downloads
Published
Issue
Section
License
Copyright (c) 2025 Dwika Ardya, Muhammad Iqbal
This work is licensed under a Creative Commons Attribution-NonCommercial-ShareAlike 4.0 International License.
International Conference On Islamic Community Studies, Education and Science, its website, and the articles published therein are licensed under a 
