GMMSampling: a new model-based, data difficulty-driven resampling method for multi-class imbalanced data.
Iwo NaglikMateusz LangoPublished in: Mach. Learn. (2024)
Keyphrases
- synthetic data
- data sets
- input data
- training data
- database
- statistical methods
- test data
- statistical analysis
- data collection
- correlation analysis
- segmentation method
- missing data
- detection method
- training samples
- preprocessing
- computational cost
- dynamic programming
- data mining techniques
- prior knowledge
- significant improvement
- image data
- prior information
- original data
- data structure
- high precision
- data processing
- data points
- probabilistic model
- high accuracy
- edge detection
- high quality
- information loss
- spectral clustering
- support vector machine
- raw data
- missing values
- labeled data
- data analysis
- pairwise
- data driven
- high dimensional data
- clustering method
- cost function