Add a patch to include missing column cluster names in predictions from music_prop, filled with 0
When the results from music_prop do not contain the same column names as unique cluster classes in the SCE object, the operations between matrices will fail because MOD0, e.g., is created using n_cells (the number of cluster classes), but mod0 is created using proportions of music_prop.
Not sure this is the way to solve it, just proposing a workaround, which I haven't fully tested.
Add a patch to include missing column cluster names in predictions from
music_prop, filled with 0When the results from music_prop do not contain the same column names as unique cluster classes in the SCE object, the operations between matrices will fail because MOD0, e.g., is created using
n_cells(the number of cluster classes), but mod0 is created using proportions of music_prop.Not sure this is the way to solve it, just proposing a workaround, which I haven't fully tested.