engleski

Fractionally integrated process with power-law correlations in variables and magnitudes

We propose a new stochastic process, de ned by only two parameters, 1 and 2 , which generates power-law correlations or anticorrelations in the variable increments, and power-law correlations in their magnitudes to account for empirical observations in a range of physical systems. To demonstrate the utility of our process, we consider the case with power-law anticorrelations in the variable increments and power-law correlations in their magnitudes to model scaling properties in heartbeat data. The parameters 1 and 2 control separately the correlations in the increments and their magnitudes, and we nd linear relations between each parameter and the corresponding correlation exponent.

stohasic process ; correlation ; heartbeat ; weather data

nije evidentirano