Clustering reduced interval data using Hausdorff distance (Q880910)

The authors consider problems of clustering and principal components analysis (PCA), as a pre-processing before clustering for multivariate interval data. Different interval PCA schemes are discussed with special attention to those using convex hulls (CH) of projected intervals. A hierarchical agglomerative clustering for CH's based on the Hausdorff distance is described. Clustering of the Breiman waves dataset is considered as an example.