Describe the feature or idea you want to propose

In /clustering/_elastic_som.py , scikit-learn has been imported only for check_random_state method.

import numpy as np
from numpy.random import RandomState
from sklearn.utils.random import check_random_state

from aeon.clustering.base import BaseClusterer
from aeon.distances import get_alignment_path_function, pairwise_distance

This adds extra scikit-learn dependency. aeon codebase generally avoids calling scikit-learn methods directly, and uses own implementation rather. So this import looks redundant.

Describe your proposed solution

Remove scikit-learn import and implement a numpy based _check_random_state method. aeon generally uses numpy-pandas methods directly in its codebase so this implementation seems reasonable.

Describe alternatives you've considered, if relevant

No response

Additional context

No response