Some wavelet analyses of point process data
Signals, Systems amp; Computers, 1997. Conference Record of the Thirty-First Asilomar Conference on, 2, page 1087-1091 vol.2. (November 1997)DOI: 10.1109/ACSSC.1997.679073
Abstract
Point processes may be described by conditional intensity functions and moments. Wavelet analysis provides a means of parameterizing such quantities in the nonstationary case. The parameters of the wavelet expansion may themselves be estimated by the method of moments or likelihood analysis. These ideas are illustrated for data sets arising from nerve cells firing and earthquakes occurring. In particular wavelet parameterized rate, autointensity and conditional intensity functions are estimated.