Abstract
An adaptive resonance theory neural network based intrusion detection approach is proposed. The approach processes both network-based and host-based data. After analyzing both the spatial and temporal associate relationship between intrusion behaviors, the associate information of the intrusion feature data is processed to detect effectively the associate relationship between intrusion behaviors. With the abilities of self-learning and self-organization, with better stability-plasticity tradeoff and the capability of quick recognition of the adaptive resonance theory, the approach can be used to detect user behaviors in real-time with good performance, especially in the recognition of unknown attacks.
Original language | English |
---|---|
Pages (from-to) | 701-704 |
Number of pages | 4 |
Journal | Beijing Ligong Daxue Xuebao/Transaction of Beijing Institute of Technology |
Volume | 24 |
Issue number | 8 |
Publication status | Published - Aug 2004 |
Keywords
- Adaptive resonance theory (ART)
- Intrusion detection
- Neural networks