Restoring balance: principled under/oversampling of data for optimal classification.
Emanuele LoffredoMauro PastoreSimona CoccoRémi MonassonPublished in: CoRR (2024)
Keyphrases
- data sets
- database
- data analysis
- data distribution
- high quality
- data collection
- data mining techniques
- training samples
- statistical analysis
- small number
- knowledge discovery
- data sources
- pattern recognition
- data structure
- feature selection
- probability distribution
- raw data
- databases
- spatial data
- missing data
- synthetic data
- training examples
- image data
- clustering algorithm
- computer systems
- data processing
- input data
- xml documents
- remote sensing
- classification algorithm
- prior knowledge
- classification rules
- dynamic programming
- data quality
- data reduction