A wavelet-domain Markov model for volatility clustering

Volatility clustering is one of the fundamental properties of volatility dynamics. However, most existing volatility clustering methods can only work with one special time-scale representation of time series, ignoring the interaction of traders with different time horizon and the information exist in multiple time-scales. In this paper, we use a wavelet-domain Markov chain model to study the volatility clustering of time series from a multi-resolution view. Experimental results on real datasets show that this method is generally effective in volatility clustering analysis. And we try to explain the result in a way consisting with the properties of the method used.