Approximation theory perspective on data processing

When
Start: 04/06/2016 - 4:15pm
End  : 04/06/2016 - 5:15pm

Category
Colloquium

Speaker
Hrushikesh Mhaskar (CGU/Caltech)

Abstract

We argue that data based classification and regression problems can be viewed profitably as problems of function approximation. We describe in detail a case study in the theory of function approximation in case of approximation of continuous 2*pi–periodic functions by trigonometric polynomials. In particular, we will discuss the equivalence theorems, and some relatively new research regarding local approximation based on Fourier information. Several extensions to other settings are mentioned, and a few applications and potential applications are discussed.

Where
Argue Auditorium, Millikan, Pomona College