Describe the workflow you want to enable

Currently the default value for LinearSVC and LinearSVR with regards to solving the problem in dual is True. This means that for users that are unfamiliar with support-vector machines and do not know the importance of this parameter they could be using te framework inefficiently.

I would propose for the default value to be None and a simple check to occur. If n_samples > n_features it's set to dual and otherwise it is set to primal.

Describe your proposed solution

This is a quick and easy fix and would only require a few lines of code in the fit method:

        if self.dual != None:
            pass
        else:
            if X.shape[0] < X.shape[1]:
                self.dual = True
            else:
                self.dual = False

Describe alternatives you've considered, if relevant

No response

Additional context

No response