Leveraging Semisupervised Hierarchical Stacking Temporal Convolutional Network for Anomaly Detection in IoT Communication
- 9 June 2020
- journal article
- research article
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Internet of Things Journal
- Vol. 8 (1) , 144-155
Abstract
The rapid development of the Internet of Things (IoT) accumulates a large number of communication records, which are utilized for anomaly detection in IoT communication. However, only a small part of these records can be labeled, which increases the difficulty in anomaly detection. This paper proposes a semi-supervised Hierarchical Stacking Temporal Convolutional Network (HS-TCN), which is the first semisupervised model for anomaly detection in IoT communication, and it can train unlabeled data based on a small number of labeled data. Furthermore, HS-TCN fully considers the features of streaming data in IoT communication and can weed out uncertain records. Finally, the experimental results demonstrate that HS-TCN promotes the performance of anomaly detection and attains better efficiency.Keywords
Funding Information
- National Natural Science Foundation of China (61702005, U1936220)
This publication has 38 references indexed in Scilit:
- Adaptive Elastic Net GMM Estimation With Many Invalid Moment Conditions: Simultaneous Model and Moment SelectionJournal of Business & Economic Statistics, 2017
- Distributed smart space orchestration2016
- Convolutional neural networks at constrained time cost2015
- Internet of Things in Industries: A SurveyIEEE Transactions on Industrial Informatics, 2014
- On the Complexity of Neural Network Classifiers: A Comparison Between Shallow and Deep ArchitecturesIEEE Transactions on Neural Networks and Learning Systems, 2014
- Speech recognition with deep recurrent neural networks2013
- Clusterer ensembleKnowledge-Based Systems, 2006
- Model selection for support vector machine classificationNeurocomputing, 2003
- Fuzzy support vector machinesIEEE Transactions on Neural Networks, 2002
- A streaming ensemble algorithm (SEA) for large-scale classification2001