Estimation of Smoothed Wavelet Spectra/Coherence in Continuous time from Multivariate Point-Processes
Observations from point-processes are often recorded in continuous time, binning these events can allow for analysis as a discrete time-series. However, often high-frequency dependency structure can be lost when performing such aggregations. To avoid this, we have developed a temporal smoothing method for the continuous wavelet transform (and corresponding wavelet spectra). The spectral estimators in this package thus operate on the continuous time event processes directly.
- Simply add the folder and its contents to your MATLAB file path
- Set up your data as a set of cells
E{1}, E{2}, ... E{p}each with a vector of points - Generate spectral estimates
S = tsWP( E, kappa, wavType ,'method','kernel');
- kappa = smoothing window scale
- wavType = Morlet, Mex (only these two are implemented currently)