MA-VAE: Multi-Head Attention-Based Variational Autoencoder Approach for Anomaly Detection in Multivariate Time-Series Applied to Automotive Endurance Powertrain Testing.
Lucas CorreiaJan-Christoph GoosPhilipp KleinThomas BäckAnna V. KononovaPublished in: IJCCI (2023)
Keyphrases
- anomaly detection
- multivariate time series
- intrusion detection
- network intrusion detection
- anomalous behavior
- network traffic
- detecting anomalies
- image segmentation
- intrusion detection system
- categorical data
- detect anomalies
- negative selection algorithm
- unsupervised learning
- preprocessing
- pattern recognition
- similarity measure
- data sets
- one class support vector machines