PDF

Description

This report explores methods for interpolating across high-dimensional data sets. We describe and evaluate algorithms designed for problems with 100 to 10,000 points in dimensions 2 through 40. Specialized algorithms that attempt to take advantage of properties of locality are shown to be less efficient than generalized algorithms such as gift-wrapping and linear programming. In addition, this report contains an accumulation of information on the properties of high-dimensional spaces.

Details

Files

Statistics

from
to
Export
Download Full History