Introduction to Statistics with Wavelets in SAS
David J. Corliss, Magnify Analytics · 2012
In statistics, wavelets are used to describe transitory periodic behavior. Wavelets’ property of cyclical behavior with a definite start and end finds applications in many different areas of statistics, such as EKG and brainwaves in biometrics and the way a commodity market responds to good or bad news. Many physical systems can be described by wavelets, with important applications used in the detection of new oil fields, the analysis of hearing and the study of vibrational events on the surface of stars. Although source code for wavelet analysis has existed in R for some time, published SAS programs have relied on the use of IML to execute R functionality within a SAS platform. This paper gives a brief introduction to wavelets, provides source code in base SAS for the time series analysis using the Morlet wavelet and reviews advanced wavelet capabilities available in the SAS / IML ® package.