A Robust AUC Maximization Framework with Simultaneous Outlier Detection and Feature Selection for Positive-Unlabeled Classification.
Ke RenHaichuan YangYu ZhaoMingshan XueHongyu MiaoShuai HuangJi LiuPublished in: CoRR (2018)
Keyphrases
- feature selection
- outlier detection
- classification accuracy
- text classification
- support vector
- model selection
- naive bayes
- credit card fraud detection
- high dimensionality
- decision trees
- training set
- machine learning
- unsupervised learning
- roc curve
- feature space
- machine learning algorithms
- high dimensional
- fraud detection
- feature subset
- learning algorithm
- class labels
- training data
- text categorization
- data streams
- supervised learning
- knowledge discovery
- active learning
- data analysis
- least squares
- information retrieval
- support vector machine
- density estimation
- probability estimation
- database