MI-MOTE: Multiple imputation-based minority oversampling technique for imbalanced and incomplete data classification.
Kyoham ShinJongmin HanSeokho KangPublished in: Inf. Sci. (2021)
Keyphrases
- multiple imputation
- incomplete data
- missing data
- statistical databases
- missing values
- class imbalance
- minority class
- learning bayesian networks
- incomplete data sets
- majority class
- bayesian networks
- em algorithm
- machine learning
- mutual information
- neural network
- classification accuracy
- class distribution
- nearest neighbour
- training data
- imbalanced datasets
- pattern recognition
- low dimensional
- imbalanced data
- feature space
- training set
- probabilistic model
- missing attribute values