Feature-level SMOTE: Augmenting fault samples in learnable feature space for imbalanced fault diagnosis of gas turbines.
Dan LiuShisheng ZhongLin LinMinghang ZhaoXuyun FuXueyun LiuPublished in: Expert Syst. Appl. (2024)
Keyphrases
- fault diagnosis
- feature level
- minority class
- feature space
- class distribution
- training samples
- class imbalance
- training set
- support vector machine
- fault detection
- gas turbine
- sampling methods
- expert systems
- feature representation
- wind turbine
- neural network
- multiple faults
- genetic algorithm
- decision level
- fusion process
- fuzzy logic
- feature set
- multi sensor
- hyperplane
- feature selection
- input space
- fault detection and diagnosis
- object level
- learning algorithm
- monitoring and fault diagnosis
- feature extraction
- training data
- feature vectors
- kernel function
- classification accuracy
- palmprint
- linear classifiers
- high dimensional
- active learning
- fault tree
- principal component analysis
- data fusion
- operating conditions
- dimensionality reduction
- data points
- cost sensitive
- training examples
- supervised learning
- chemical process
- fault detection and isolation
- multi sensor information fusion
- fusion scheme
- low dimensional
- wavelet transform
- multi class
- decision trees