Summary

International Symposium on Nonlinear Theory and Its Applications

2015

Session Number:A3L-A

Session:

Number:A3L-A-1

Improvements to Local Projective Noise Reduction Through Higher Order and Multiscale Refinements

Jack Murdock Moore,  Ali Karrech,  Michael Small,  

pp.113-116

Publication Date:2015/12/1

Online ISSN:2188-5079

DOI:10.34385/proc.47.A3L-A-1

PDF download (256.7KB)

Summary:
Kantz and Schreiber (2004) established a second order, single scale refinement of the first order, single scale local projective filter of Cawley and Hsu (1992); Sauer (1992). Their method better accommodates the geometry of a strange attractor underlying data. This work extends the refinement to higher order and multiple scales, thereby providing an analytic motivation for the original refinement. Additionally, statistical analysis explains the original refinement's high efficacy. The novel refinements identified in this work can achieve more effective noise reduction than established local projective filters.