Empirical error-confidence Curves for Neural Network and Gaussian Classifiers.
Gregory J. WolffDavid G. StorkArt B. OwenPublished in: Int. J. Neural Syst. (1996)
Keyphrases
- neural network
- bias variance decomposition
- training data
- multi layer perceptron
- pattern recognition
- artificial neural networks
- nearest neighbour
- support vector
- expected error
- small sample
- decision trees
- neural network model
- error rate
- svm classifier
- machine learning
- nonlinear functions
- bias variance
- training process
- recurrent neural networks
- theoretical analysis
- training set
- gaussian kernels
- bp neural network
- naive bayes
- covariance matrices
- confidence measures
- variance reduction
- class conditional
- confidence measure
- maximum likelihood
- training samples
- neural network is trained
- back propagation
- genetic algorithm
- akaike information criterion
- taylor series expansion
- semi supervised
- feature selection
- confidence levels
- classification algorithm
- decision boundary
- linear classifiers
- test set
- b spline
- gaussian mixture model