Back to index

Linear Clustering of Objects with Multiple Attributes

H.V. Jagadish, Bell Labs

Summary by Armando Fox and Steve Gribble

One-line summary: Survey of various schemes for mapping multidimensional data into a linear space, with investigation of locality-preserving properties for column-based and range queries.

Overview/Main Points


Concrete and practical "goodness metrics" for locality ("clustering") in database sense (i.e. "few disk I/O's"), and of how to preserve locality inherent in a multiple-dimension representation while mapping down to a smaller number of dimensions.


Back to index