Efficient identification of regular expressions from representative examples