Novel Meta-Features for Automated Machine Learning Model Selection in Anomaly Detection.
Milos KotlarMarija PuntZaharije RadivojevicMilos CvetanovicVeljko MilutinovicPublished in: IEEE Access (2021)
Keyphrases
- model selection
- anomaly detection
- machine learning
- unsupervised learning
- cross validation
- intrusion detection
- bayesian learning
- network anomaly detection
- hyperparameters
- detecting anomalies
- network intrusion detection
- statistical inference
- feature selection
- anomalous behavior
- intrusion detection system
- network traffic
- feature extraction
- feature set
- model selection criteria
- statistical learning
- detect anomalies
- machine learning algorithms
- one class support vector machines
- automatic model selection
- classification accuracy
- computer vision
- marginal likelihood
- feature space
- data analysis
- bayesian networks
- similarity measure
- clustering algorithm
- negative selection algorithm
- learning algorithm