2006/04/26 by Robert W. Johnson, Johnson, Robert W.
Computer Science · Mathematics · #Astrophysics (astro-ph) #Computational Physics and Python Applications #Data Analysis #FOS: Physical sciences #Image and Signal Denoising Methods #Space Physics (physics.space-ph) #Statistical and numerical algorithms #Statistics and Probability (physics.data-an)
paper · pdf · doi:10.48550/arxiv.physics/0604211
openalex publication_date 2006/04/26 · openalex created_date 2025/10/10 · openalex updated_date 2026/07/28
The traditional continuous wavelet transform is plagued by the cone-of-influence, ie wavelets which extend past either end of a finite timeseries return transform coefficients which tend to decrease as more of the wavelet is truncated. These coefficients may be corrected simply by rescaling the remaining wavelet. The corrected wavelet transform displays no cone-of-influence and maintains reconstruction as either edge is approached. As an application and example, we present the corrected wavelet transform of the (derectified) yearly International Sunspot Number, Ri, as a measure of solar magnetic activity, and compare the yearly solar magnetic power with Oerlemans' glacial global temperature reconstruction.