For paper submission, I have recently spent some time struggling to find appropriate classification. The main question is: who needs this information and why?
At first I thought that they could be used by the editor to find an appropriate editor or referees, but I have had two experiences contradicting this hypothesis. In the first one, I was only asked to provide classification AFTER the review process. In the second, I was requested to point myself to appropriate editors and referees.
In my experience, I do not look at those numbers, and that is true for other colleagues. The only exception is when I am trying to find appropriate classification and look at other related papers for inspiration. In such occasion, I sometimes found classifications which did not seem to match the content of the paper. Apparently people do not care much for this. Another aspect of the question is: how bad is it to have a bad classification?