Combining cross-validation and confidence to measure fitness.
D. Randall WilsonTony R. MartinezPublished in: IJCNN (1999)
Keyphrases
- cross validation
- model selection
- hyperparameters
- confidence measure
- support vector
- variable selection
- regression problems
- training set
- error estimates
- cross validated
- generalization error
- genetic algorithm
- genetic programming
- evolutionary algorithm
- correlation coefficient
- regularization parameter
- fitness function
- similarity measure
- feature selection
- nearest neighbor classifiers
- neural network
- population size
- unseen data
- learning algorithm