Describes opportunities for utilizing robust graph representations of data with machine learning algorithms. The authors have selected the domain of web content mining, which involves the clustering and classification of web documents based on their textual substance.