How is clustering affected by high-dimensional data, and how can the quality of clusters generated be improved in such cases?