The Development of a Firewall Application Using Deep Learning Methods for Detection of Cyber Attacks on Web Applications
Yıl 2025,
Cilt: 16 Sayı: 2, 331 - 344, 30.06.2025
Şengül Bayrak
,
Alper Karaca
,
Ferhat Toson
Mehmet Emin Tayfur
Selçuk Yavaş
Öz
The security of web applications is critical for protecting user data and taking precautions against cyber-attacks. Firewall applications are used to protect Internet users' sensitive data, take precautions against cyber-attacks and improve the user experience. Although traditional firewall approaches try to detect attacks based on certain rules and patterns, they may be insufficient against evolving and complex attack types. In this paper, we propose a deep learning based approach to detect cyber attacks on web applications. The proposed method is developed using the “FWAF” dataset and goes through stages that are data preprocessing, feature extraction, and data scaling. Six different deep learning models are evaluated to determine the method with the highest accuracy and reliability. The results obtained played an important role in the selection of the optimum method by comparing the model performances. The selected model was put live on the Jetson Nano platform to test its real-time operability and its performance was evaluated. Experimental results show that deep learning-based approaches offer higher accuracy rates compared to traditional methods and can adapt to new types of attacks. This study makes an important contribution to the field of cyber security by providing an effective and dynamic security solution for detecting attacks against web applications.
Proje Numarası
İstanbul Sabahattin Zaim Üniversitesi BAP-2023-36 ve TUBITAK 2209 - 1919B012322054
Kaynakça
- [1] A. Brown, M. Gupta, and M. Abdelsalam, “Automated machine learning for deep learning based malware detection,” Computers & Security, vol. 137, p. 103582, Nov. 2023, doi: 10.1016/j.cose.2023.103582.
- [2] A. Paul, V. Sharma, and O. Olukoya, “SQL injection attack: Detection, prioritization & prevention,” Journal of Information Security and Applications, vol. 85, p. 103871, Aug. 2024, doi: 10.1016/j.jisa.2024.103871.
- [3] B. Dawadi, B. Adhikari, and D. Srivastava, “Deep learning technique-enabled web application firewall for the detection of web attacks,” Sensors, vol. 23, no. 4, p. 2073, Feb. 2023, doi: 10.3390/s23042073.
- [4] K. Köksal, B. Doğan, and Z. A. Altıkardeş, “Topluluk sınıflandırıcı kullanarak zararlı url tespiti,” Veri Bilimi, vol. 4, no. 3, p. 113-122, 2021.
- [5] E. Uçar, M. Ucar, and M. O. İncetaş, “A Deep learning approach for detection of malicious URLs,” In 6th International Management Information Systems Conference, p. 12-20, October, 2019.
- [6] F. D. Abdi, and L. Wenjuan, “Malicious URL detection using convolutional neural network,” Journal International Journal of Computer Science, Engineering and Information Technology, vol. 7, no. 6, p. 1-8, 2017.
- [7] F. Tiryaki, Ü. Şentürk, and İ. Yücedağ, “Developing and evaluating an artificial intelligence model for malicious url detection,” Avrupa Bilim ve Teknoloji Dergisi, vol.47, p.13-17, 2023.
- [8] M. Moghimi and A. Y. Varjani, “New rule-based phishing detection method,” Expert Systems With Applications, vol. 53, pp. 231–242, Jan. 2016, doi: 10.1016/j.eswa.2016.01.028.
- [9] M. Ismail, S. Alrabaee, K.-K. R. Choo, L. Ali, and S. Harous, “A comprehensive evaluation of machine learning algorithms for web application attack detection with knowledge graph integration,” Mobile Networks and Applications, Jul. 2024, doi: 10.1007/s11036-024-02367-z.
- [10] Ş. Bayrak, and B. Ardanuç, “Web uygulamaları için derin öğrenme yöntemiyle güvenlik duvarı uygulaması firewall application with the deep learning method for the web applications,” Sinyal İşlemeleri Uygulamaları Konferansı, 1-5, 2022.
- [11] K. Baysal, “Derin öğrenme temelli yeni nesil güvenlik duvarının tasarlanması”, Trakya Üniversitesi, Fen Bilimleri Enstitüsü, Doktora Tezi, 2022.
- [12] S. Toprak, and A. Yavuz, “Web application firewall based on anomaly detection using deep learning,” Acta Infologica, vol. 0, no. 0, p. 0, Oct. 2022, doi: 10.26650/acin.1039042.
- [13] S. Demir, Z. Aslan, “K-NN, NN ve feature selection yöntemleri ile firewall verilerinin sınıflandırması”, Anadolu Bil Meslek Yüksekokulu Dergisi, vol. 17, no.66, pp.139-148, 2023.
- [14] F. Demi̇r, “Siber saldırı tespiti için makine öğrenmesi yöntemlerinin performanslarının incelenmesi,” Balıkesir Üniversitesi Fen Bilimleri Enstitüsü Dergisi, vol. 23, no. 2, pp. 782–791, May 2021, doi: 10.25092/baunfbed.876338.
- [15] M. Takaoğlu and Ç. Özer, “The effect of machine learning on intruson detectıon systems,” Uluslararası Yönetim Bilişim Sistemleri ve Bilgisayar Bilimleri Dergisi, vol. 3, no. 1, pp. 11–22, Jun. 2019, doi: 10.33461/uybisbbd.558192.
- [16] Einy, S. Makina öğrenmesile biyometrik sahtekarlığa ve ağ anormallik tespitine dayalı saldırı tespiti. Sakarya Üniversitesi, 2021.
- [17] Y. E. Seyyar, A. G. Yavuz, and H. M. Unver, “Detection of web attacks using the BERT model,” 2022 30th Signal Processing and Communications Applications Conference (SIU), pp. 1–4, May 2022, doi: 10.1109/siu55565.2022.9864721.
- [18] E. Baş, and A. A. Süzen, Web uygulama zafiyetlerinin keşfinde açıklanabilir yapay zekânın yeri. All Science Proceeding, 2023.
- [19] I. Abasıkeleş-Turgut, and R. Daş, “Anomaly and intrusion detection systems for smart grids,” Elsevier, 2024, pp. 231–270. doi: https://doi.org/10.1016/B978-0-443-14066-2.00005-0.
- [20] R. S. Arslan, “FastTrafficAnalyzer: An efficient method for intrusion detection systems to analyze network traffic,” DÜMF Mühendislik Dergisi, pp. 565–572, Sep. 2021, https://doi.org/10.24012/dumf.1001881.
- [21] Y. Xu, Q. Zhang, H. Deng, Z. Liu, C. Yang, and Y. Fang, “Unknown web attack threat detection based on large language model,” Applied Soft Computing, pp. 112905–112905,Feb.2025, https://doi.org/10.1016/j.asoc.2025.112905.
- [22] Osama Salahat, Y. Assaf, and A. Younis, “Securing web applications using machine learning,” Najah.edu, Feb.03,2025. https://repository.najah.edu/items/97bd9567-c9cd-4722-914e-4fba60c321c8
- [23] T. Wang, Y. Xu, and Z. Tang, “Toward fast network intrusion detection for web services: partial-flow feature extraction and dataset construction,” International Journal of Web Information Systems, Dec. 2024, doi: https://doi.org/10.1108/ijwis-09-2024-0261.
- [24] R. Gallego, L. Pérez, Marcos Luengo Viñuela, and María Concepción Vega-Hernández, “Artificial intelligent web application firewall for advanced detection of web injection attacks,” Expert Systems, Nov. 2023, doi: https://doi.org/10.1111/exsy.13505.
- [25] L. Hickman, S. Thapa, L. Tay, M. Cao, and P. Srinivasan, “Text preprocessing for text mining in organizational research: review and recommendations,” Organizational Research Methods, vol. 25, no. 1, pp. 114–146, Nov. 2020, doi: 10.1177/1094428120971683.
- [26] C. P. Chai, “Comparison of text preprocessing methods,” Natural Language Engineering, vol. 29, no. 3, pp. 509–553, Jun. 2022, doi: 10.1017/s1351324922000213.
- [27] S. Feuerriegel et al., “Using natural language processing to analyse text data in behavioural science,” Nature Reviews Psychology, Jan. 2025, doi: 10.1038/s44159-024-00392-z.
- [28] M. Ahsan, M. Mahmud, P. Saha, K. Gupta, and Z. Siddique, “Effect of data scaling methods on machine learning algorithms and model performance,” Technologies, vol. 9, no. 3, p. 52, Jul. 2021, doi: 10.3390/technologies9030052.
- [29] J. Barbero-Gómez, R. P. M. Cruz, J. S. Cardoso, P. A. Gutiérrez, and C. Hervás-Martínez, “CNN explanation methods for ordinal regression tasks,” Neurocomputing, p. 128878, Nov. 2024, doi: 10.1016/j.neucom.2024.128878.
- [30] S. Shao, “Enhancing sentiment analysis with a CNN-Stacked LSTM hybrid model,” ITM Web of Conferences, vol. 70, p. 02002, Jan. 2025, doi: 10.1051/itmconf/20257002002.
- [31] B. Ticona, F. Carranza, and V. Cotik, “Indigenous languages spoken in Argentina: a survey of NLP and speech resources,” arXiv (Cornell University), Jan. 2025, doi: 10.48550/arxiv.2501.09943.
- [32] A. Dayan and A. Yilmaz, “Modelling the machines’ language with natural language processing and machine learning algorithms,” DÜMF Mühendislik Dergisi, Jul. 2022, doi: 10.24012/dumf.1131565.
- [33] S. Bayrak, “Unveiling intrusions: explainable SVM approaches for addressing encrypted Wi-Fi traffic in UAV networks,” Knowledge and Information Systems, vol. 66, no. 11, pp. 6675–6695, Jul. 2024, doi: https://doi.org/10.1007/s10115-024-02181-9.
- [34] A. Almsaeed, (2018). AdminLTE control panel template. URl: https://almsaeedstudio. com/(visited on 17/4/2023).
- [35] F. Ahmad, “GitHub - faizann24/Fwaf-Machine-Learning-driven-Web-Application-Firewall: Machine learning driven web application firewall to detect malicious queries with high accuracy.,” GitHub. https://github.com/faizann24/Fwaf-Machine-Learning-driven-Web-Application Firewall?tab=readme-ov-file]
- [36] P. Tufekci and E. Uzun, “Author detection by using different term weighting schemes,” 2013 21st Signal Processing and Communications Applications Conference (SIU), pp. 1–4, Apr. 2013, doi: https://doi.org/10.1109/siu.2013.6531190.
Web Uygulamalarına Yönelik Siber Saldırıların Tespitinde Derin Öğrenme Yöntemleri Kullanılarak Güvenlik Duvarı Uygulamasının Geliştirilmesi
Yıl 2025,
Cilt: 16 Sayı: 2, 331 - 344, 30.06.2025
Şengül Bayrak
,
Alper Karaca
,
Ferhat Toson
Mehmet Emin Tayfur
Selçuk Yavaş
Öz
Web uygulamalarının güvenliği, kullanıcı verilerinin korunması ve siber saldırılara karşı önlem alınması açısından kritik bir öneme sahiptir. İnternet kullanıcılarının hassas verilerini koruma, siber saldırılara karşı önlem alma ve kullanıcı deneyimini iyileştirme gibi amaçlarla güvenlik duvarı uygulamaları kullanılmaktadır. Geleneksel güvenlik duvarı yaklaşımları, belirli kurallar ve kalıplara dayanarak saldırıları tespit etmeye çalışsa da gelişen ve karmaşık hale gelen saldırı türlerine karşı yetersiz kalabilmektedir. Bu çalışmada, web uygulamalarına yönelik siber saldırıları tespit etmek için derin öğrenme tabanlı bir yaklaşım önerilmektedir. Önerilen yöntem, “FWAF” veri seti kullanılarak geliştirilmiş ve veri ön işleme, özellik çıkarımı ve veri ölçekleme aşamalarından geçirilmiştir. Altı farklı derin öğrenme modeli değerlendirilerek en yüksek doğruluk ve güvenilirliği sağlayan yöntem belirlenmiştir. Elde edilen sonuçlar, model performanslarını karşılaştırmalı olarak inceleyerek optimum yöntemin seçiminde önemli bir rol oynamıştır. Seçilen model, gerçek zamanlı çalışabilirliğini test etmek amacıyla Jetson Nano platformu üzerinde canlıya alınarak performansı değerlendirilmiştir. Deneysel sonuçlar, derin öğrenme tabanlı yaklaşımların geleneksel yöntemlere kıyasla daha yüksek doğruluk oranları sunduğunu ve yeni saldırı türlerine adapte olabildiğini göstermektedir. Bu çalışma, web uygulamalarına yönelik saldırıların tespitinde etkili ve dinamik bir güvenlik çözümü sunarak, siber güvenlik alanında önemli bir katkı sağlamaktadır.
Etik Beyan
Hazırlanan makalede etik kurul izni alınmasına gerek yoktur.
Hazırlanan makalede herhangi bir kişi/kurum ile çıkar çatışması bulunmamaktadır.
Destekleyen Kurum
İstanbul Sabahattin Zaim Üniversitesi ve TÜBİTAK
Proje Numarası
İstanbul Sabahattin Zaim Üniversitesi BAP-2023-36 ve TUBITAK 2209 - 1919B012322054
Teşekkür
Bu proje İstanbul Sabahattin Zaim Üniversitesi BAP-2023-36 numaralı proje ile desteklenmiştir.
Ayrıca bu çalışma TUBITAK 2209 - 1919B012322054 kodlu proje ile desteklenmiştir.
Kaynakça
- [1] A. Brown, M. Gupta, and M. Abdelsalam, “Automated machine learning for deep learning based malware detection,” Computers & Security, vol. 137, p. 103582, Nov. 2023, doi: 10.1016/j.cose.2023.103582.
- [2] A. Paul, V. Sharma, and O. Olukoya, “SQL injection attack: Detection, prioritization & prevention,” Journal of Information Security and Applications, vol. 85, p. 103871, Aug. 2024, doi: 10.1016/j.jisa.2024.103871.
- [3] B. Dawadi, B. Adhikari, and D. Srivastava, “Deep learning technique-enabled web application firewall for the detection of web attacks,” Sensors, vol. 23, no. 4, p. 2073, Feb. 2023, doi: 10.3390/s23042073.
- [4] K. Köksal, B. Doğan, and Z. A. Altıkardeş, “Topluluk sınıflandırıcı kullanarak zararlı url tespiti,” Veri Bilimi, vol. 4, no. 3, p. 113-122, 2021.
- [5] E. Uçar, M. Ucar, and M. O. İncetaş, “A Deep learning approach for detection of malicious URLs,” In 6th International Management Information Systems Conference, p. 12-20, October, 2019.
- [6] F. D. Abdi, and L. Wenjuan, “Malicious URL detection using convolutional neural network,” Journal International Journal of Computer Science, Engineering and Information Technology, vol. 7, no. 6, p. 1-8, 2017.
- [7] F. Tiryaki, Ü. Şentürk, and İ. Yücedağ, “Developing and evaluating an artificial intelligence model for malicious url detection,” Avrupa Bilim ve Teknoloji Dergisi, vol.47, p.13-17, 2023.
- [8] M. Moghimi and A. Y. Varjani, “New rule-based phishing detection method,” Expert Systems With Applications, vol. 53, pp. 231–242, Jan. 2016, doi: 10.1016/j.eswa.2016.01.028.
- [9] M. Ismail, S. Alrabaee, K.-K. R. Choo, L. Ali, and S. Harous, “A comprehensive evaluation of machine learning algorithms for web application attack detection with knowledge graph integration,” Mobile Networks and Applications, Jul. 2024, doi: 10.1007/s11036-024-02367-z.
- [10] Ş. Bayrak, and B. Ardanuç, “Web uygulamaları için derin öğrenme yöntemiyle güvenlik duvarı uygulaması firewall application with the deep learning method for the web applications,” Sinyal İşlemeleri Uygulamaları Konferansı, 1-5, 2022.
- [11] K. Baysal, “Derin öğrenme temelli yeni nesil güvenlik duvarının tasarlanması”, Trakya Üniversitesi, Fen Bilimleri Enstitüsü, Doktora Tezi, 2022.
- [12] S. Toprak, and A. Yavuz, “Web application firewall based on anomaly detection using deep learning,” Acta Infologica, vol. 0, no. 0, p. 0, Oct. 2022, doi: 10.26650/acin.1039042.
- [13] S. Demir, Z. Aslan, “K-NN, NN ve feature selection yöntemleri ile firewall verilerinin sınıflandırması”, Anadolu Bil Meslek Yüksekokulu Dergisi, vol. 17, no.66, pp.139-148, 2023.
- [14] F. Demi̇r, “Siber saldırı tespiti için makine öğrenmesi yöntemlerinin performanslarının incelenmesi,” Balıkesir Üniversitesi Fen Bilimleri Enstitüsü Dergisi, vol. 23, no. 2, pp. 782–791, May 2021, doi: 10.25092/baunfbed.876338.
- [15] M. Takaoğlu and Ç. Özer, “The effect of machine learning on intruson detectıon systems,” Uluslararası Yönetim Bilişim Sistemleri ve Bilgisayar Bilimleri Dergisi, vol. 3, no. 1, pp. 11–22, Jun. 2019, doi: 10.33461/uybisbbd.558192.
- [16] Einy, S. Makina öğrenmesile biyometrik sahtekarlığa ve ağ anormallik tespitine dayalı saldırı tespiti. Sakarya Üniversitesi, 2021.
- [17] Y. E. Seyyar, A. G. Yavuz, and H. M. Unver, “Detection of web attacks using the BERT model,” 2022 30th Signal Processing and Communications Applications Conference (SIU), pp. 1–4, May 2022, doi: 10.1109/siu55565.2022.9864721.
- [18] E. Baş, and A. A. Süzen, Web uygulama zafiyetlerinin keşfinde açıklanabilir yapay zekânın yeri. All Science Proceeding, 2023.
- [19] I. Abasıkeleş-Turgut, and R. Daş, “Anomaly and intrusion detection systems for smart grids,” Elsevier, 2024, pp. 231–270. doi: https://doi.org/10.1016/B978-0-443-14066-2.00005-0.
- [20] R. S. Arslan, “FastTrafficAnalyzer: An efficient method for intrusion detection systems to analyze network traffic,” DÜMF Mühendislik Dergisi, pp. 565–572, Sep. 2021, https://doi.org/10.24012/dumf.1001881.
- [21] Y. Xu, Q. Zhang, H. Deng, Z. Liu, C. Yang, and Y. Fang, “Unknown web attack threat detection based on large language model,” Applied Soft Computing, pp. 112905–112905,Feb.2025, https://doi.org/10.1016/j.asoc.2025.112905.
- [22] Osama Salahat, Y. Assaf, and A. Younis, “Securing web applications using machine learning,” Najah.edu, Feb.03,2025. https://repository.najah.edu/items/97bd9567-c9cd-4722-914e-4fba60c321c8
- [23] T. Wang, Y. Xu, and Z. Tang, “Toward fast network intrusion detection for web services: partial-flow feature extraction and dataset construction,” International Journal of Web Information Systems, Dec. 2024, doi: https://doi.org/10.1108/ijwis-09-2024-0261.
- [24] R. Gallego, L. Pérez, Marcos Luengo Viñuela, and María Concepción Vega-Hernández, “Artificial intelligent web application firewall for advanced detection of web injection attacks,” Expert Systems, Nov. 2023, doi: https://doi.org/10.1111/exsy.13505.
- [25] L. Hickman, S. Thapa, L. Tay, M. Cao, and P. Srinivasan, “Text preprocessing for text mining in organizational research: review and recommendations,” Organizational Research Methods, vol. 25, no. 1, pp. 114–146, Nov. 2020, doi: 10.1177/1094428120971683.
- [26] C. P. Chai, “Comparison of text preprocessing methods,” Natural Language Engineering, vol. 29, no. 3, pp. 509–553, Jun. 2022, doi: 10.1017/s1351324922000213.
- [27] S. Feuerriegel et al., “Using natural language processing to analyse text data in behavioural science,” Nature Reviews Psychology, Jan. 2025, doi: 10.1038/s44159-024-00392-z.
- [28] M. Ahsan, M. Mahmud, P. Saha, K. Gupta, and Z. Siddique, “Effect of data scaling methods on machine learning algorithms and model performance,” Technologies, vol. 9, no. 3, p. 52, Jul. 2021, doi: 10.3390/technologies9030052.
- [29] J. Barbero-Gómez, R. P. M. Cruz, J. S. Cardoso, P. A. Gutiérrez, and C. Hervás-Martínez, “CNN explanation methods for ordinal regression tasks,” Neurocomputing, p. 128878, Nov. 2024, doi: 10.1016/j.neucom.2024.128878.
- [30] S. Shao, “Enhancing sentiment analysis with a CNN-Stacked LSTM hybrid model,” ITM Web of Conferences, vol. 70, p. 02002, Jan. 2025, doi: 10.1051/itmconf/20257002002.
- [31] B. Ticona, F. Carranza, and V. Cotik, “Indigenous languages spoken in Argentina: a survey of NLP and speech resources,” arXiv (Cornell University), Jan. 2025, doi: 10.48550/arxiv.2501.09943.
- [32] A. Dayan and A. Yilmaz, “Modelling the machines’ language with natural language processing and machine learning algorithms,” DÜMF Mühendislik Dergisi, Jul. 2022, doi: 10.24012/dumf.1131565.
- [33] S. Bayrak, “Unveiling intrusions: explainable SVM approaches for addressing encrypted Wi-Fi traffic in UAV networks,” Knowledge and Information Systems, vol. 66, no. 11, pp. 6675–6695, Jul. 2024, doi: https://doi.org/10.1007/s10115-024-02181-9.
- [34] A. Almsaeed, (2018). AdminLTE control panel template. URl: https://almsaeedstudio. com/(visited on 17/4/2023).
- [35] F. Ahmad, “GitHub - faizann24/Fwaf-Machine-Learning-driven-Web-Application-Firewall: Machine learning driven web application firewall to detect malicious queries with high accuracy.,” GitHub. https://github.com/faizann24/Fwaf-Machine-Learning-driven-Web-Application Firewall?tab=readme-ov-file]
- [36] P. Tufekci and E. Uzun, “Author detection by using different term weighting schemes,” 2013 21st Signal Processing and Communications Applications Conference (SIU), pp. 1–4, Apr. 2013, doi: https://doi.org/10.1109/siu.2013.6531190.