Unsupervised Learning of Distributional Properties can Supplement Human Labeling and Increase Active Learning Efficiency in Anomaly Detection.
Jaturong KongmaneeMark H. ChignellKhilan JerathAbhay RamanPublished in: CoRR (2023)
Keyphrases
- anomaly detection
- unsupervised learning
- active learning
- supervised learning
- semi supervised
- detecting anomalies
- intrusion detection
- anomalous behavior
- network traffic
- machine learning
- unsupervised anomaly detection
- detecting anomalous
- expectation maximization
- computer security
- network intrusion detection
- self organizing maps
- dimensionality reduction
- behavior analysis
- feature selection
- intrusion detection system
- malware detection
- active learning strategies
- object recognition
- cumulative sum
- labeled data
- network anomaly detection
- data mining
- named entities
- learning process
- genetic algorithm
- network intrusion
- detect anomalies
- learning algorithm
- training set
- association rules
- network security
- human subjects