Warder: Online Insider Threat Detection System Using Multi-Feature Modeling and Graph-Based Correlation
The DOI matches its Crossref record; 15 of the 15 references it cites were found in a registry or an index.
Abstract
Existing insider threat detection models and frameworks generally focus on characterizing and detecting malicious insiders, for example by fusing behavioral analysis, machine learning, psychological characters, management measures, etc. However, it remains challenging to design a practical insider threat detection scheme that can be efficiently implemented and deployed in a real-world system. For example, existing approaches focus on extracting features from user behavioral activities but they lack in-depth correlation and decision making for suspected alerts; thus, resulting in high false positives and low detection accuracy. In this work, we propose a novel online insider threat detection system, Warder, which leverages diverse feature dimensions (using neural language processing) and fuses content and behavior features to create a user's daily profile to facilitate threat detection. Besides, hypergraph-based threat scenario feature tree is designed to correlate suspicious users' activities with threat scenarios to further screen the users. In practice, Warder can also be constantly updated using newly discovered features and threat scenarios. We evaluate the performance of Warder using the public CMU CERT dataset, as well as that of approaches from the Oxford group and CMU group. Findings from the evaluation demonstrate that Warder outperforms the other two competing approaches.
Overview
Integrity signals
Registry
Cites
Statistics
Verification
Claims the paper makes
Claims other papers make about this one
Each is checked against this paper’s own text: does it say what the citing paper says it does?
In the citing paper
“Moreover, supervised learning methods, such as support vector machines (SVM) and random forests, leverage labeled datasets to train models capable of discriminating between normal and malicious activities [73], [74], [75].”
From A Review of Recent Advances, Challenges, and Opportunities in Malicious Insider Threat Detection Using Machine Learning Methods (2024) · cites this paper as [75]
Citations
Cites
Bridging the Gap: A Pragmatic Approach to Generating Insider Threat Data
Glasser, J., & Lindauer, B. (2013) · 2013 IEEE Security and Privacy Workshops · doi:10.1109/spw.2013.37 · Cited by 344
Automated Insider Threat Detection System Using User and Role-Based Profile Assessment
Legg, P. A., Buckley, O., Goldsmith, M., & Creese, S. (2017) · IEEE Systems Journal · doi:10.1109/jsyst.2015.2438442 · Cited by 145 · cited 4×
“These models generally focus on building comprehensive profile of users and then using anomaly detection [2], [3], scenario analysis [4] and graph analysis [5] to detect malicious insiders.”
InfluentialFeature Correlation Hypergraph: Exploiting High-order Potentials for Multimodal Recognition
Zhang, L., Gao, Y., Hong, C., Feng, Y., Zhu, J., & Cai, D. (2014) · IEEE Transactions on Cybernetics · doi:10.1109/tcyb.2013.2285219 · Cited by 139 · cited 1×
“Hyper graph [13] is a graph where an edge can connect any number of vertices.”
BackgroundBehavioral Analysis of Insider Threat: A Survey and Bootstrapped Prediction in Imbalanced Data
Azaria, A., Richardson, A., Kraus, S., & Subrahmanian, V. S. (2014) · IEEE Transactions on Computational Social Systems · doi:10.1109/tcss.2014.2377811 · Cited by 133 · cited 2×
“…could be overcame by the decision module, which suggests that our designed decision layer can greatly enhance the precision and recall rate of threat detection precision. c) Performance comparison: We compare Warder with two representative approaches of the Oxford group [2] and CMU group [14].”
MethodScenario-Based Insider Threat Detection From Cyber Activities
Chattopadhyay, P., Wang, L., & Tan, Y.-P. (2018) · IEEE Transactions on Computational Social Systems · doi:10.1109/tcss.2018.2857473 · Cited by 115 · cited 1×
“In [10], for example, the authors proposed a system which designed a time window for correlation and threat decision.”
BackgroundBF-IoT: Securing the IoT Networks via Fingerprinting-Based Device Authentication
Gu, T., & Mohapatra, P. (2018) · 2018 IEEE 15th International Conference on Mobile Ad Hoc and Sensor Systems (MASS) · doi:10.1109/mass.2018.00047 · Cited by 49 · cited 1×
“These models generally focus on building comprehensive profile of users and then using anomaly detection [2], [3], scenario analysis [4] and graph analysis [5] to detect malicious insiders.”
Self-citationTowards Insider Threat Detection Using Psychophysiological Signals
Hashem, Y., Takabi, H., GhasemiGol, M., & Dantu, R. (2015) · Proceedings of the 7th ACM CCS International Workshop on Managing Insider Security Threats · doi:10.1145/2808783.2808792 · Cited by 35 · cited 1×
“There are also some competing approaches on insider threat detection, such as those in [15], [16].”
BackgroundSOFIT: Sociotechnical and Organizational Factors for Insider Threat
Greitzer, F., Purl, J., Leong, Y. M., & Becker, D. E. S. (2018) · 2018 IEEE Security and Privacy Workshops (SPW) · doi:10.1109/spw.2018.00035 · Cited by 32 · cited 1×
“Although both objective and subjective factors [6] of users have been utilized to build comprehensive frameworks, these models are not easily trained and evaluated due to lack of realworld data.”
Kandias, M., Gritzalis, D. A., Stavrou, V., & Nikoloulis, K. (2016) · Computers & Security · doi:10.1016/j.cose.2016.12.003 · Cited by 29 · cited 1×
“There are also some competing approaches on insider threat detection, such as those in [15], [16].”
BackgroundMylrea, M., Gourisetti, S. N. G., Larimer, C., & Noonan, C. F. (2018) · IEEE Security and Privacy Workshops (SPW) · doi:10.1109/spw.2018.00036 · Cited by 23 · cited 1×
“Therefore, behavioral activities of users are widely used to characterize and build profiles of users, as evidenced by existing research [7].”
Background
Research works printed but not linked: 1
- schulze, 0, Insider Threat 2018 Report
Cited by
Alzaabi, F. R., & Mehmood, A. (2024) · IEEE Access · doi:10.1109/access.2024.3369906 · Cited by 204
“Moreover, supervised learning methods, such as support vector machines (SVM) and random forests, leverage labeled datasets to train models capable of discriminating between normal and malicious activities [73], [74], [75].”
MethodMetamorphic Malware and Obfuscation: A Survey of Techniques, Variants, and Generation Kits
Brezinski, K., & Ferens, K. (2023) · Security and Communication Networks · doi:10.1155/2023/8227751 · Cited by 80
IoTGaze: IoT Security Enforcement via Wireless Context Analysis
Gu, T., Fang, Z., Abhishek, A., Fu, H., Hu, P., & Mohapatra, P. (2020) · IEEE Conference on Computer Communications (INFOCOM) · doi:10.1109/infocom41043.2020.9155459 · Cited by 64
“Graph-based detection approaches [32], [33] can also be applied to IoT to detect anomalies.”
MethodSelf-citationITDBERT: Temporal-semantic Representation for Insider Threat Detection
Huang, W., Zhu, H., Li, C., Lv, Q., Wang, Y., & Yang, H. (2021) · 2021 IEEE Symposium on Computers and Communications (ISCC) · doi:10.1109/iscc53001.2021.9631538 · Cited by 37
Self-citationWei, Y., Chow, K.-P., & Yiu, S.-M. (2021) · Forensic Science International: Digital Investigation · doi:10.1016/j.fsidi.2021.301126 · Cited by 32
“[13] proposed to use a graph-based correlation model to do a second step decision making, its actual implementation is through hand crafted construction of the scenario feature trees and doing a simple matching on the recognized anomalous activities in the previous step, which is not applicable in the real case with l…”
BackgroundSITD: Insider Threat Detection Using Siamese Architecture on Imbalanced Data
Zhou, S., Wang, L., Yang, J., & Zhan, P. (2022) · International Conference on Computer Supported Cooperative Work in Design (CSCWD) · doi:10.1109/cscwd54268.2022.9776049 · Cited by 18
What Do We Know About the Psychology of Insider Threats?
Ruohonen, J., & Saddiqa, M. (2025) · Lecture notes of the Institute for Computer Sciences, Social Informatics and Telecommunications Engineering · doi:10.1007/978-3-031-89363-6_11 · Cited by 13
“Emails [44, 45, 54, 60, 85], visited web pages [44, 45, 54], and social media [65] have provided the typical sources for empirical data.”
BackgroundZhang, F., Ma, X., & Huang, W. (2022) · International Joint Conference on Neural Networks (IJCNN) · doi:10.1109/ijcnn55064.2022.9892163 · Cited by 12
“1) Insider Threat Detection: TYPical insider threat detec tion methods extract features of user's behavior and process them by machine learning models [5], [6], [16], [17].”
MethodSelf-citationA systematic review on insider threat detection using natural language processing
Kundiya, K., & Haribhakta, Y. (2025) · International Journal of Information Security · doi:10.1007/s10207-025-01145-6 · Cited by 9
“NLP techniques like NER rationales for tackling issues like data imbalance and linguistic ambiguity in cybersecurity contexts [19, 65 ].”
A Proposed Approach to Analyze Insider Threat Detection Using Emails
Mittal, A., & Garg, U. (2022) · 2022 3rd International Conference on Computation, Automation and Knowledge Management (ICCAKM) · doi:10.1109/iccakm54721.2022.9990361 · Cited by 8
UAG: User Action Graph Based on System Logs for Insider Threat Detection
Wang, X., Jiang, J., Wang, Y., Lv, Q., & Wang, L. (2023) · 2023 IEEE Symposium on Computers and Communications (ISCC) · doi:10.1109/iscc58397.2023.10218139 · Cited by 6
“Typical insider threat detection methods extract features of user’s behavior and process them by machine learning models [11] [12] [6].”
BackgroundSelf-citationGu, H., He, D., & Zhou, X. S. (2026) · ACM SIGKDD Conference on Knowledge Discovery and Data Mining (KDD) · doi:10.1145/3770854.3780237 · Cited by 1
“MAGNN [5] has demonstrated effective predictions by aggregating information from heterogeneous datasets.”
Taming Toxic Talk: Detecting Offensive Online Content in HausaPreprint
Adam, F. M., Zandam, A. Y., & Inuwa-Dutse, I. (2024) · Research Square · doi:10.21203/rs.3.rs-4266465/v1
Reviewing Graph Construction Rules in Graph-Based Insider Threat Detection Models
Chung, J. Z. A., Yong, K. S. C., Han, J., & Then, P. (2025) · IEEE Symposium on Computer Applications & Industrial Electronics (ISCAIE) · doi:10.1109/iscaie64985.2025.11081151
“Other approaches, such as those in [8-10], mitigate this issue by condensing nodes into user-days rather than users.”
Outlier Explanations in Data Streams - Applications for Environmental Data
Teixeira, C. F. (2024) · Portuguese National Funding Agency for Science, Research and Technology (RCAAP Project by FCT) · doi:10.34626/4w7a-rr74
Studying Edge Relationship in Homogeneous Graph-based Insider Threat Detection
Chung, J. Z. A., Yong, K. S. C., Han, J., & Then, P. (2026) · International Journal of Sustainable Construction Engineering and Technology (Universiti Tun Hussein Onn Malaysia)
Metadata
- Authors
- Jiang, J., Chen, J., Gu, T., Choo, K.-K. R., Liu, C., Yu, M., Huang, W., & Mohapatra, P.
- Affiliations
- Chinese Academy of Sciences (CN)
- University of California, Davis (US)
- The University of Texas at San Antonio (US)
- Venue
- MILCOM 2019 - 2019 IEEE Military Communications Conference (MILCOM)
- Publisher
- IEEE
- Published
- 2019-11
- Pages
- 1-6
- OpenAlex
- W3011330084
- Cited by
- 14
- References
- 15
- Citations by year
- 2026: 1 · 2025: 3 · 2024: 2 · 2023: 2 · 2022: 3 · 2021: 2 · 2020: 1
- FWCI
- 1.19
- Type
- conference paper
- Language
- en
- Open access
- closed
- Licence
- Locations
- IEEE Military Communications Conference (MILCOM) (publishedVersion)
- Topics
- Physical Sciences › Computer Science › Information Systems › Information and Cyber Security
- Physical Sciences › Computer Science › Computer Networks and Communications › Network Security and Intrusion Detection
- Physical Sciences › Computer Science › Information Systems › User Authentication and Security Systems
- Keywords
- insider threat detection, false positive reduction
About these checks
Registry, retraction, Cites and statistics results are computed, not judged, with one exception: a printed reference that no index lists is read once by a model before it counts as a citation error. Claims are extracted by a model and kept only when their quote appears verbatim in the text; each cited source is then checked, with a model’s help, for whether it says what the claim relies on. Those verdicts can be wrong — read the claim’s quote and the cited source before relying on one.
Statistics and arithmetic are recomputed, not judged.
Confidence bands are shown only for fields calibrated on our annotated test set. None is calibrated yet, so no band is shown.
What counts as a claim ›How we measure confidence ›How we check this ›
Are you an author, or do you see a problem with this page? Request a review
Data sources
Bibliographic data: Semantic Scholar, OpenAlex, Crossref.
Contains information from the Semantic Scholar Academic Graph Datasets, made available under the ODC Attribution License (ODC-BY 1.0).
Semantic Scholar’s data is described in: Kinney, R., et al. (2023). The Semantic Scholar Open Data Platform. arXiv:2301.10140.