Interesting patterns for clustering high-dimensional data