Nonparametric Data Selection for Improvement of Parametric Neural Learning: A Cumulant-Surrogate Method.
Gustavo DecoBernd SchürmannPublished in: ICANN (1996)
Keyphrases
- synthetic data
- significant improvement
- statistical methods
- input data
- test data
- data sets
- prior knowledge
- data analysis
- neural learning
- database
- neural network
- data sources
- noisy data
- image data
- clustering method
- prior information
- segmentation method
- missing data
- objective function
- data collection
- data mining techniques
- training samples
- feature extraction
- similarity measure
- decision trees
- clustering algorithm
- knowledge discovery
- probability distribution
- incremental learning
- bayesian methods
- training data