Core-Preserving Algorithms

We define a class of algorithms for constructing coresets of (geometric) data sets, and show that algorithms in this class can be dynamized efficiently in the insertiononly (data stream) model. As a result, we show that for a set of points in fixed dimensions, additive and multiplicative -coresets for the k-center problem can be maintained in O(1) and O(k) time respectively, using a data structure whose size is independent of the size of the input. We also provide a faster streaming algorithm for maintaining -coresets for fat extent-related problems such as diameter and minimum enclosing ball.

• Type: Conference
• Year: 2008
• Where: CCCG
• Authors: Hamid Zarrabi-Zadeh
• Keywords: Algorithm, CCCG 2008, Computational Geometry, Data Sets, Data Stream