Boosting training for PDF malware classifier via active learning.
Yuanzhang LiXinxin WangZhiwei ShiRuyun ZhangJingfeng XueZhi WangPublished in: Int. J. Intell. Syst. (2022)
Keyphrases
- active learning
- training examples
- training set
- partially labeled data
- rare classes
- boosted classifiers
- learning algorithm
- training process
- training samples
- supervised learning
- discriminative classifiers
- early stopping
- query by committee
- batch mode
- weak classifiers
- training error
- adaboost algorithm
- semi supervised
- generalization error
- ground truth labels
- feature selection
- training data
- labeled instances
- annotation effort
- active learner
- probability density function
- decision stumps
- label noise
- weak learners
- labeling effort
- cost sensitive
- ensemble learning
- random sampling
- test set
- multiple classifier systems
- decision trees
- machine learning
- semi supervised learning
- labeled data
- ensemble methods
- base classifiers
- class labels
- training dataset
- svm classifier
- support vector machine
- classification accuracy
- feature space
- sample selection
- support vector
- rare class
- feature set
- classification algorithm
- linear classifiers
- object detection
- face detection
- class imbalance
- improving classification accuracy
- co training
- multi class
- base learners
- ensemble classifier
- text classifiers