Improving the representation of legal case texts with information extraction methods