Estimating MLP Generalisation Ability without a Test Set using Fast, Approximate Leave-one-out Cross-validation.
Andrew J. MylesAlan F. MurrayA. Robin WallaceJohn BarnardGordon SmithPublished in: Neural Comput. Appl. (1997)
Keyphrases
- test set
- leave one out cross validation
- confidence intervals
- error rate
- training set
- k nearest neighbour
- test data
- model selection
- multilayer perceptron
- neural network
- multi layer perceptron
- support vector machine
- cross validation
- variable selection
- training data
- generalization error
- radial basis function
- feature subset
- feature extraction
- image processing
- back propagation
- multi class
- artificial neural networks