Deploying Isolation Forest at the Edge: A Synthetic Da-ta-driven Approach for Real-time IoT Anomaly Detection

References

[1] Ahmad Z, Khan A S, Nisar K, Haider I, Hassan R, Haque M R, Tarmizi S, & Rodrigues J J P C. Anomaly detection using deep neural network for IoT architecture. Applied Sciences, 2021, 11(15): 7050. DOI: 10.3390/app11157050. [journal paper]

[2] Aluwala A. AI-driven anomaly detection in network monitoring techniques and tools. Journal of Artificial Intelligence & Cloud Computing, 2024, SRC/JAICC-329. DOI: 10.47363/JAICC/2024(3)310. [journal paper]

[3] Aswad F M, Ahmed A M S, Alhammadi N, & Khalaf B A. Deep learning in distributed denial-of-service attacks detection method for Internet of Things networks. Journal of Intelligent Systems, 2023, 32(1): 20220155. DOI: 10.1515/jisys-2022-0155. [journal paper]

[4] Choi W H, & Kim J. Unsupervised learning approach for anomaly detection in industrial control systems. Applied System Innovation, 2024, 7(2): 18. DOI: 10.3390/asi7020018. [journal paper]

[5] Chicco D, Oneto L, & Tavazzi E. Eleven quick tips for data cleaning and feature engineering. PLOS Computational Biology, 2022, 18(12): e1010718. DOI: 10.1371/journal.pcbi.1010718. [journal paper]

[6] Hosseinzadeh M, Rahmani A M, Vo B, Bidaki M, Masdari M, & Zangakani, M. Improving security using SVM-based anomaly detection: Issues and challenges. Soft Computing, 2021, 25(4): 3195-3223. DOI: 10.1007/s00500-020-05373-x. [journal paper]

[7] Karra A E. The effect of using data pre-processing by imputations in handling missing values. Indonesian Journal of Electrical Engineering and Informatics, 2022, 10(2): 375-384. DOI: 10.52549/ijeei.v10i2.3730. [journal paper]

[8] Kaur N. Robustness and security in deep learning: Adversarial attacks and countermeasures. Journal of Electrical Systems, 2024, 20(3s): 1250-1257. DOI: 10.52783/jes.1436. [journal paper]

[9] Khan M A, Khan M A, Jan S U, Ahmad J, Jamal S S, Shah A A, Pitropakis N, & Buchanan W J. A deep learning-based intrusion detection system for MQTT enabled IoT. Sensors, 2021, 21(21): 7016. DOI: 10.3390/s21217016. [journal paper]

[10] Markevych M, & Dawson M. A review of enhancing intrusion detection systems for cybersecurity using artificial intelligence (AI). International Conference Knowledge-Based Organization, Jun. 2023, 29(3). DOI: 10.2478/kbo-2023-0072. [conference paper]

[11] Olateju O O, Okon S U, Igwenagu U, & Salami A. Combating the challenges of false positives in AI-driven anomaly detection systems and enhancing data security in the cloud. Asian Journal of Research in Computer Science, 2024, 17(6): 264-292. DOI: 10.9734/ajrcos/2024/v17i6472. [journal paper]

[12] Omol E, Mburu L, Abuonji P, & Onyango D. Anomaly detection in IoT sensor data using machine learning techniques for predictive maintenance in smart grids. International Journal of Science Technology & Management, 2024, 5(1): 201-210. DOI: 10.46729/ijstm.v5i1.1028. [journal paper]

[13] Sharma A, & Babbar H. Machine learning-based anomaly detection in the Internet of Things. 3rd Asian Conference on Innovation in Technology (ASIANCON), Aug. 2023, pp. 1-6. DOI: 10.1109/ASIANCON58793.2023.10270100. [conference paper]

[14] Singh R, & Gill S S. Edge AI: A survey. Internet of Things and Cyber-Physical Systems, 2023, 3: 71-92. DOI: 10.1016/j.iotcps.2023.02.004. [thesis]

@article{1706999,
      author = {Judah Idowu},
      title = {Deploying Isolation Forest at the Edge: A Synthetic Da-ta-driven Approach for Real-time IoT Anomaly Detection},
      journal = {Iconic Research And Engineering Journals},
      year = {2025},
      volume = {8},
      number = {7},
      pages = {643-652},
      issn = {2456-8880},
      url = {https://www.irejournals.com/formatedpaper/1706999.pdf},
      abstract = {This research addresses the challenge of detecting network anomalies in IoT environments by developing and evaluating AI-driven methodologies, specifically deploying Isolation Forest at the edge. A key obstacle in this domain is the scarcity of labeled datasets representing diverse anomaly types. To overcome this, we employed a synthetic data-driven approach, using statistical modeling to generate datasets that mimic real-world IoT scenarios, encompassing a wide range of normal and anomalous behaviors. This synthetic data enabled robust training and evaluation of our anomaly detection model. Performance was assessed using precision, recall, F1-score, and AUC-ROC. While the model demonstrated high preci-sion for normal data points, its performance in identifying anomalous data points was limited, as reflected in the low recall, F1-score, and ROC-AUC values. Despite these limitations, the study demonstrates the potential of Isolation Forest for detecting simpler anomalies like outliers and drifts, contributing to reduced false positives. Furthermore, the research emulated edge computing principles?distributed processing and reduced latency?to facilitate real-time analysis, which is crucial for mitigating security breaches and system failures in dynamic IoT environments. Although the simulation did not involve physical edge devices, it successfully demonstrated the feasibility of deploying Isolation Forest at the edge for real-time anomaly detection, highlighting its potential for enhancing operational efficiency, risk management, and security in sectors like healthcare, where early anomaly detection is critical.},
      keywords = {Artificial Intelligence, Edge Computing, IoT Security, Machine Learning, Statistical Methods},
      month = {January},
  }