Checking for presence of features in fit()

When checking for the present of the `features` in the fit method, this raises an error unless the type is a list (for example using a numpy array). This is because the check for its presence is being done by way of an `if features`: shouldn't this be `if features is not None` for a more robust implementation?