Wavelet filtering for prediction in time series analysis
Tommaso Minerva · IRIS UNIMORE (University of Modena and Reggio Emilia) · 2010
In this paper I describe a wavelet filtering approach to separate a time series, the signal, into its main components. With this approach I can separate stochastic from structural components. The statistical predictive analysis will be performed on the filtered signal while the stochastic term could be a-posteriori reintroduced through statistical simulation approaches (such as Markov Chain Monte Carlo). The proposed metodology has been applied to financial time series to predict both returns and risk.