[BUG] fix MeanScale.mean ignoring the scale parameter

[VividhDesign]

Reference Issues/PRs

Fixes #1185

What does this implement/fix? Explain your changes.

MeanScale._mean now returns mu + sigma * d.mean() instead of mu + d.mean().

Does your contribution introduce a new dependency? If yes, which one?

No.

What should a reviewer concentrate their feedback on?

One-line change in meanscale.py.

Did you add any tests for the change?

Yes, skpro/distributions/tests/test_meanscale.py:

• the mean and var of the docstring example against closed form;
• the mean against ppf(0.5) for a symmetric component, with array mu and sigma.

Both fail on main and pass with the fix.

Any other comments?

(AI-assisted; reviewed and tested by me.)

PR checklist

• I've added myself to the list of contributors (bug, code)
• The PR title starts with [BUG]