Illuminating the black box: Explaining machine learning algorithms that detect fake news on Twitter