Published Paper PDF: View PDF
DOI: https://doi.org/10.63345/ijrsml.v10.i4.1
Confirmation Letter: View
Certificate: View Certificate
Ishu Anand Jaiswal
Independent Researcher
Civil Lines, , Kanpur, UP, India-208001
Abstract— NLP has become a critical enabler in understanding and operationalizing textual security artefacts; however, current research remains fragmented between policy-focused and log-focused methodologies. On one hand, existing studies provide strong foundations for extracting access-control rules, assessing ambiguity, and analyzing completeness in natural-language security policies. On the other hand, parallel work demonstrates the efficacy of NLP-driven feature learning and deep sequence models for anomaly detection in system logs. These strands seldom intersect, leaving a substantive research gap: the absence of integrated frameworks that align high-level policy intent with the low-level system behavior captured in logs. This gap constrains security teams from automating compliance verification, detecting policy-violating activities, and obtaining interpretable, end-to-end visibility across security controls. The paper bridges this gap by proposing an NLP-driven architecture that jointly models security policies and system logs, allowing automatic linking of policy clauses and operational evidence. Our approach incorporates semantic role extraction, linguistic ambiguity scoring, log template mapping, and contextual sequence modeling to realize a unified representation space for policy statements and log events. By correlating these representations, the system allows for automated compliance checks, interpretable anomaly detection, and natural language querying of policies and logs. Experimental evaluation based on real-world datasets demonstrates improved coverage of policy-to-log traceability, reduced false positives in log anomaly detection, and enhanced analyst trust due to explainable outputs.
Keywords— Natural Language Processing, Security Policies, Log Analysis, Anomaly Detection, Compliance Automation
References
[1] S. Wilson et al., “The Creation and Analysis of a Website Privacy Policy Corpus,” Proc. 54th Annual Meeting of the Association for Computational Linguistics, 2016.
[2] J. R. Reidenberg et al., “Ambiguity in Privacy Policies and the Impact of Regulation,” Journal of Legal Studies, vol. 45, no. S2, pp. S163–S190, 2016.
[3] J. Bhatia and T. D. Breaux, “Towards an Information Type Lexicon for Privacy Policies,” Proc. 8th IEEE Int. Workshop on Requirements Engineering and Law, pp. 19–24, 2015.
[4] J. Bhatia and T. D. Breaux, “Semantic Incompleteness in Privacy Policy Goals,” Proc. 26th IEEE Int. Requirements Engineering Conf., pp. 159–169, 2018.
[5] J. M. Del Álamo et al., “A Systematic Mapping Study on Automated Analysis of Privacy Policy Texts,” Computing, vol. 104, pp. 2979–3010, 2022.
[6] X. Xiao, A. Paradkar, S. Thummalapenta, and T. Xie, “Automated Extraction of Security Policies from Natural-Language Software Documents,” Proc. ACM SIGSOFT Int. Symp. Foundations of Software Engineering (FSE), 2012.
[7] N. Papanikolaou, “Natural Language Processing of Rules and Regulations for Privacy and Security in Cloud Computing,” in Trust, Privacy and Security in Digital Business, Springer, pp. 124–136, 2012.
[8] M. Alohaly, K. Elmiligi, and M. Elrabaa, “Automated Extraction of Attributes from Natural Language Access Control Policies,” Cybersecurity, vol. 2, no. 7, 2019.
[9] P. Shi et al., “Checking Network Security Policy Violations via Natural Language Queries,” Proc. 2021 IEEE Conf. on Communications and Network Security, 2021.
[10] H. Lundblad, “An NLP Approach to Assess Information Security Policies,” M.Sc. thesis, Chalmers Univ. of Technology, 2022.
[11] J. Wang et al., “LogEvent2Vec: LogEvent-to-Vector Based Anomaly Detection for Large-Scale Logs in Internet of Things,” Sensors, vol. 20, no. 9, 2451, 2020.
[12] P. Ryciak, K. Wasielewska, and A. Janicki, “Anomaly Detection in Log Files Using Selected Natural Language Processing Methods,” Applied Sciences, vol. 12, no. 10, 5089, 2022.
[13] M. Landauer et al., “Deep Learning for Anomaly Detection in Log Data: A Survey,” arXiv preprint arXiv:2207.03820, 2022.
[14] C. Almodovar et al., “Can Language Models Help in System Security?” Proc. ALTA Workshop, 2022