Synthetic Samples Generation for Imbalance Class Distribution with LSTM Recurrent Neural Networks.
Biprodip PalArnab Kanti TarafderMd. Shahinur RahmanPublished in: ICCA (2020)
Keyphrases
- recurrent neural networks
- class distribution
- training samples
- training set
- minority class
- long short term memory
- class imbalance
- majority class
- imbalanced datasets
- training data
- cost sensitive
- misclassification costs
- test set
- neural network
- feed forward
- training examples
- recurrent networks
- imbalanced data
- reservoir computing
- test data
- highly skewed
- data sets
- echo state networks
- highly imbalanced
- sampling methods
- imbalanced data sets
- unlabeled data
- cost sensitive learning
- base classifiers
- nonlinear dynamic systems
- active learning
- class labels
- error rate
- artificial neural networks
- semi supervised
- feature space
- supervised learning
- prior knowledge
- nearest neighbour
- support vector machine
- multi class
- classification error