Is feature selection secure against training data poisoning?
Huang XiaoBattista BiggioGavin BrownGiorgio FumeraClaudia EckertFabio RoliPublished in: CoRR (2018)
Keyphrases
- training data
- feature selection
- classification accuracy
- classification models
- text categorization
- training set
- naive bayes
- support vector machine
- test data
- learning algorithm
- data sets
- decision trees
- information gain
- feature set
- training process
- support vector
- multi class
- prior knowledge
- feature selection algorithms
- selected features
- mutual information
- text classification
- training dataset
- method for feature selection
- security mechanisms
- feature subset
- training samples
- feature extraction
- feature space
- supervised learning
- machine learning
- unsupervised feature selection
- training instances
- security requirements
- security issues
- dimensionality reduction
- generalization error
- semi supervised learning
- microarray data
- high dimensionality
- knn
- unsupervised learning
- labeled data
- discriminative features
- key management
- standard model
- authentication protocol
- class labels