A wavelet exploration of the BVL index

Wavelets allow for a more flexible characterization of time series than both spectral and classical time series methods, by representing them with basis functions that separate between time and scale. In this paper, we briefly present the the main definitions and results of both wavelet and wavelet packet analysis and apply some of them to the Lisbon stock exchange index (IBVL). We denoise the series, separate trend and cycle, and present a Monte carlo estimation of the fractal dimension of the series