Similarity Measures for Ordered Histograms