nmf_imaging: Second Release

Enabling Data Imputation in Non-negative Matrix Factorization (NMF) Target Modeling

• The nmf_imaging.NMFmodelling function now can take a mask_data_imputation input, which is a two dimensional binary mask that matches the size of the trg, and 0 means you will impute the values for the corresponding matrix elements. However, if some of the 0's are not covered in your NMF components, the corresponding element values cannot be imputed.

Allowing Missing Data in NMF Component Calculation

• The nmf_imaging.NMFcomponents function's mask input now allows a three dimensional binary mask that matches the size of the refs input, and 0 means that you will ignore the the corresponding element's input value.