Analisis Komparatif Model Unsupervised Learning (Isolation Forest Vs. Autoencoder) Untuk Deteksi Anomali Pada Klaim Asuransi Kesehatan
DOI:
https://doi.org/10.31849/q5st4j47Keywords:
Unsupervised Learning, Isolation Forest, Autoencoder, Anomaly Detection, Health Insurance ClaimsAbstract
Insurance is a financial product that serves to protect customers from unexpected future risks. One of its main products is health insurance, which plays a vital role in covering high medical expenses, especially for individuals who lack sufficient financial resources. Anomaly detection in health insurance claims is a crucial step in preventing fraud and maintaining the operational efficiency of insurance companies. This study compares two unsupervised Learning models, namely Isolation Forest and Autoencoder, in detecting anomalies in health insurance claim data. The evaluation is conducted using metrics such as AUC, accuracy, precision, recall, and F1-score. The results show that the Isolation Forest model performs more consistently and excels in terms of accuracy and F1-score compared to the Autoencoder. Meanwhile, the Autoencoder demonstrates a higher recall in detecting anomalous claims, but its low precision reduces its overall detection effectiveness. Based on these findings, Isolation Forest is recommended as a more reliable method for anomaly detection in this case.
References
[1] Q. yumansyah Qori, A. Turmudi Zy, and M. Fatchan, “Prediksi Jumlah Kasus Klaim Indemnity Dengan Menggunakan Algoritma Regresi Linear Pada Asuransi Mandiri Inhealth,” Bull. Inf. Technol., vol. 4, no. 3, pp. 299–305, 2023, doi: 10.47065/bit.v4i3.733.
[2] T. K. Mackey, K. Miyachi, D. Fung, S. Qian, and J. Short, “Combating health care fraud and abuse: Conceptualization and prototyping study of a blockchain antifraud framework,” J. Med. Internet Res., vol. 22, no. 9, 2020, doi: 10.2196/18623.
[3] N. Kumaraswamy, M. K. Markey, T. Ekin, J. C. Barner, and K. Rascati, “Healthcare Fraud Data Mining Methods: A Look Back and Look Ahead,” Perspect. Heal. Inf. Manag., vol. 19, no. 1, 2022.
[4] T. Azad and P. William, “Fraud detection in healthcare billing and claims Fraud detection in healthcare billing and claims,” Int. J. Sci. Res. Arch., vol. 13, no. 2, 2024, doi: 10.30574/ijsra.2024.13.2.2606.
[5] D. E. Prasetyo, G. A. A. Wulandari, Z. Meini, and Fauziah, “Identifikasi Fraud Dalam Pemeriksaan Internal Melalui Data Analytics,” J. EQUITY, vol. 26, no. 1, pp. 78–98, 2024, doi: 10.34209/equ.v26i1.7418.
[6] P. K. Baruah et al., “Integrating Machine Learning Models with Business Rule Triggers to Boost Performance in Health Insurance Fraud Detection : A Case Study,” Var. Adv. Sci. Risk, vol. 18, 2025.
[7] W. Hilal, S. A. Gadsden, and J. Yawney, “Financial Fraud: A Review of Anomaly Detection Techniques and Recent Advances,” Expert Syst. Appl., vol. 193, p. 116429, 2022, doi: 10.1016/j.eswa.2021.116429.
[8] G. S. W. Prabuningrat, D. P. Hostiadi, and N. L. P. Srinadi, “Klasifikasi Deteksi Anomali Menggunakan Metode Machine Learning,” Semin. Has. Penelit. Inform. dan Komput. 2024, vol. 1, no. 2, p. 2024, 2024, [Online]. Available: http://www.jubikom.unpak.ac.id
[9] D. B. Santoso and Y. Wahyuni, “Sestem Log Web Server Sebagai Pendeteksi Anomali Menggunakan Isolation Forest,” J. Apl. Bisnis dan Komput., vol. 4, no. 3, pp. 2807–5986, 2024, [Online]. Available: http://www.jubikom.unpak.ac.id
[10] M. F. Dzulqarnain, A. Fadlil, and I. Riadi, “Implementasi CNN berbasis Autoencoder pada Klasifikasi Pola Batik,” Proceeding Informatics Collab. Dessimenation Meet. Infocoding 2025, pp. 189–194, 2025.
[11] A. Zulfikar, F. A. Rahmani, and N. Azizah, “Deteksi Anomali Menggunakan Isolation Forest Belanja Barang Persediaan Konsumsi Pada Satuan Kerja Kepolisian Republik Indonesia,” J. Manaj. Perbendaharaan, vol. 4, no. 1, pp. 1–15, 2023, doi: 10.33105/jmp.v4i1.435.
[12] R. Ardiansyah, L. Sunardi, P. S. Informatika, U. B. Insan, K. Jaringan, and I. Forest, “Implementasi Metode Isolation Forest Untuk Deteksi Anomali Dalam Data Jaringan,” Semin. Nas. ESCAF (Economic, Soc. Sci. Comput. Agric. Fish. 4th, pp. 933–939, 2025.
[13] S. Nazir, S. Patel, and D. Patel, “Autoencoder Based Anomaly Detection for SCADA Networks,” Int. J. Artif. Intell. Mach. Learn., vol. 11, no. 2, pp. 83–99, 2021, doi: 10.4018/ijaiml.20210701.oa6.
[14] H. Torabi, S. L. Mirtaheri, and S. Greco, “Practical autoencoder based anomaly detection by using vector reconstruction error,” Cybersecurity, vol. 6, no. 1, pp. 1–13, 2023, doi: 10.1186/s42400-022-00134-9.
[15] D. R. K. Saputra, Y. V. Via, and A. N. Sihananto, “Deteksi Anomali Menggunakan Ensemble Learning Dan Random Oversampling Pada Penipuan Transaksi Keuangan,” J. Inform. dan Tek. Elektro Terap., vol. 12, no. 3, pp. 2779–2788, 2024, doi: 10.23960/jitet.v12i3.4910.
[16] A. Jayadie et al., Pembiayaan Kesehatan, no. September. Bandung: MEDIA SAINS INDONESIA, 2023.
[17] R. P. Dewa and Windarto, “Deteksi Anomali Jaringan Menggunakan Isolation Forest pada Log Wazuh dengan Pemberitahuan WhatsApp di PT XYZ Network Anomaly Detection Using Isolation Forest on Wazuh Logs with WhatsApp Notifications at PT XYZ,” KRESNA J. Ris. dan Pengabdi. Masy., vol. 4, pp. 208–216, 2024.
[18] F. S. Nugraha and H. F. Pardede, “Autoencoder untuk Sistem Prediksi Berat Lahir Bayi,” J. Teknol. Inf. dan Ilmu Komput., vol. 9, no. 2, pp. 235–244, 2022, doi: 10.25126/jtiik.2022923868.
[19] I. Rianto and I. P. I. Santosa, Data Preparation untuk Machine Learning & Deep Learning. Penerbit Andi, 2025.
[20] W. R. Nurputra, W. Witanti, and A. Komarudin, “Principal Component Analysis (PCA) Untuk Meningkatkan Hasil Klasterisasi Penjualan Video Game Menggunakan Algoritma K-Means,” J. LOCUS Penelit. Pengabdi., vol. 4, no. 8, pp. 4683–4703, 2025.
