How limited training data can allow a neural network to outperform an 'optimal' statistical classifier.
Les T. NilesHarvey F. SilvermanGary N. TajchmanMarcia A. BushPublished in: ICASSP (1989)
Keyphrases
- training data
- neural network
- binary decision tree
- training process
- learning algorithm
- fuzzy artmap
- decision trees
- training instances
- support vector machine
- training set
- training samples
- training patterns
- test data
- class labels
- unseen data
- backpropagation neural network
- test set
- training examples
- dynamic programming
- manually labeled training data
- statistical analysis
- minimum error
- neural network model
- class distribution
- artificial neural networks
- labeled data
- feature values
- learning vector quantization
- prior knowledge
- text classifiers
- genetic algorithm
- pattern recognition
- feature space
- training dataset
- classification accuracy
- classification models
- semi supervised
- input data
- data sets