Is Feature Selection Secure against Training Data Poisoning?

Huang Xiao Battista Biggio Gavin Brown Giorgio Fumera Claudia Eckert Fabio Roli

Published in: ICML (2015)

Keyphrases

training data
feature selection
classification accuracy
classification models
text categorization
naive bayes
test data
feature space
training set
support vector machine
feature set
decision trees

learning algorithm
data sets
mutual information
training process
key management
feature extraction
machine learning
supervised learning
labeled data
security issues
text classification
dimensionality reduction

training samples
training instances
test set
prior knowledge
security requirements
redundant features
cryptographic protocols
irrelevant features
support vector
security analysis
feature weighting
training examples

model selection
training dataset
feature selection algorithms
semi supervised learning
feature subset
lightweight
method for feature selection
sensitive data
learned from training data
unsupervised learning
class labels