Addressing the Overlapping Data Problem in Classification Using the One-vs-One Decomposition Strategy.
José A. SáezMikel GalarBartosz KrawczykPublished in: IEEE Access (2019)
Keyphrases
- data sets
- data processing
- data reduction
- missing data
- synthetic data
- data collection
- small number
- data structure
- training data
- database
- original data
- raw data
- experimental data
- classification algorithm
- sensor data
- machine learning
- computer systems
- learning algorithm
- feature selection
- knowledge discovery
- probability distribution
- data analysis
- high quality
- decision trees
- machine learning algorithms
- multi class
- databases
- end users
- statistical methods
- preprocessing
- data quality
- extracted features