Regularization in Relevance Learning Vector Quantization Using l one Norms
This work addresses the need for sparse feature selection in hyperspectral data analysis to improve classification by fading down unnecessary spectral bands, representing an incremental advancement in regularization techniques for LVQ.
The paper tackles the problem of achieving sparse relevance profiles in prototype-based relevance learning vector quantization (LVQ) for hyperspectral data classification by proposing a method using l1 regularization via LASSO optimization, resulting in a gradient learning scheme with a differentiable approximation of the l1-norm that has an upper error bound and is extended to matrix learning variants.
We propose in this contribution a method for l one regularization in prototype based relevance learning vector quantization (LVQ) for sparse relevance profiles. Sparse relevance profiles in hyperspectral data analysis fade down those spectral bands which are not necessary for classification. In particular, we consider the sparsity in the relevance profile enforced by LASSO optimization. The latter one is obtained by a gradient learning scheme using a differentiable parametrized approximation of the $l_{1}$-norm, which has an upper error bound. We extend this regularization idea also to the matrix learning variant of LVQ as the natural generalization of relevance learning.