Anomaly Detection in IOT Data

This paper presented a unified IoT anomaly detection framework using statistical, machine learning, and deep learning techniques. The system was implemented as a Flask-based application that supports data upload, preprocessing, anomaly detection, visualization, and model comparison. The use of an industrial IoT dataset and the inclusion of metrics such as accuracy, precision, recall, and F1-score make the work more aligned with research expectations.[1][2] The dashboard results show 3931 total anomalies, with the statistical model selected as the best model for the given run. The framework achieves accuracy of 0.072, precision of 1.0, recall of 0.072, and F1-score of 0.1343. Although the recall is low, the system successfully demonstrates comparative anomaly detection across three model families and provides a practical basis for future improvement.[1] Future work can focus on real-time streaming data, more advanced benchmark datasets, explainability, and stronger deep learning or hybrid architectures. Adding ROC-AUC, PR-AUC, and a full ablation experiment with exact numbers.

P. Malhotra et al., "LSTM-based encoder-decoder for multi-sensor anomaly detection," in ICML Workshop on Anomaly Detection, 2016.[1] [2] F. T. Liu et al., "Isolation Forest," in IEEE ICDM, pp. 413-422, 2008.[1] [3] S. Hochreiter and J. Schmidhuber, "Long Short-Term Memory," Neural Comput., vol. 9, no. 8, pp. 1735-1780, 1997.[1] [4] V. Chandola et al., "Anomaly Detection: A Survey," ACM Comput. Surv., vol. 41, no. 3, pp. 1-58, 2009.[1] [5] R. Chalapathy and S. Chawla, "Deep Learning for Anomaly Detection," ACM Comput. Surv., vol. 52, no. 2, pp. 1-38, 2019.[1] [6] Md. Sad Abdullah Sami, Mushfiquzzaman Abid, “Unsupervised Anomaly Detection for Smart IoT Devices: Performance and Resource Comparison” IEEE 2nd International Conference on Computing, Applications and Systems (COMPAS 2025) 23–24 October 2025, [7] Mohamed Bachar, Azeddine Khiat, Kamal El Guemmat, Hybrid Autoencoder and Isolation Forest for IoT Anomaly Detection with a Novel Model , Vol. 16, No. 1, 2026, 31123-31129 [8] Shibzan Shahanas , Afnaj Akthar , Saanna Anand , Rakshitha , Dr. Amirthavalli.M