SIMPLE_MLP performs a one-hidden-layer neural network classification
Syntax
CLASS = SIMPLE_MLP(LEARNSET,TESTSET,ALGO,NBEPOCHS,GOAL)
Description
CLASS = SIMPLE_MLP(LEARNSET,TESTSET,ALGO,NBEPOCHS,GOAL)
LEARNSET, TESTSET are 'dataset' objects.
GOAL : first stopping condition, the global error rate. Common value is 0.01.
NBEPOCHS : second stopping condition, the number of epochs. Common value is 50.
ALGO :
- 'trainlm' Levenberg-Marquardt, slow and memory consuming but efficient.
- 'trainrp' Resilient backpropagation, fast, NBEPOCHS bigger.
- 'trainbfg' BFGS Quasi-Newton, fast, NBEPOCHS bigger.
'Neural Networks' Toolbox required.
See Also
BAG_MLP, BAGFS_MLP, MFS_MLP, simple_knn, simple_c45, simple_lda
[ Help Desk ]