In x-ray spectromicroscopy, a set of images can be acquired across an absorption edge to reveal chemical speciation. We previously described the use of non-negative matrix approximation methods for improved classification and analysis of these types of data. We present here an approach to find appropriate values of regularization parameters for this optimization approach.
Keywords: multivariate statistical analysis; non-negative matrix analysis; optimization; x-ray microscopy; x-ray spectromicroscopy.