BUG : SOMModel now herits from MachineLearningModel<TInput,...> instead of MachineLearningModel<itk:variablelengthvector<TInput>,...>

DR is now based on Machine Learning Models, it is now possible to reduce de dimension of the pca model (it has to be changed in PCAModel.txx for now

Sparse ae added, parameters can now be defined for each layers (noise, regularization, rho and beta)

Shark pca and autoencoders DoPredict's method (ie without using batches) is now working, it shouldn't be used thought, as using batches is more efficient with Shark