Smoothest Local Interpolation Formulas for Equally Spaced Data.
Abstract
Let a moving-average interpolation formula for equally spaced data, exact for the degree r, have a basic function L e C to the m-1 power of finite support with L to the (m) power piecewise continuous. Such a formula is called 'smoothest' when the integral of the square of L to the (m) power over the support of L is smallest. If m, r, and the support of L are given, either there is no such formula or there is a unique smoothest formula, for which L is a piecewise polynomial of degree at least r and at most max(r, 2m - 1), uniquely characterized by certain conditions on the location of its knots and the jumps occurring there. A similar result is obtained if consideration is limited to formulas that preserve (i.e., do not smooth) the given data.
Document Details
- Document Type
- Technical Report
- Publication Date
- Aug 01, 1981
- Accession Number
- ADA103860
Entities
People
- Hubert Vaughan
- Thomas N.E. Greville
Organizations
- University of Wisconsin–Madison