I try to understand the 2nd formula stated in the picture. It yields the intensity/volatility of an fGN process. It depends solely on H ?? Why is that?
How is this volatility different from simply calculating standard deviation of an empirical or simulated record?
And why is it different from ARMA filters, who operate on some driving variance of arbitrary intensity? Does the self similarity law impose this?
Our ultimate goal is upscaling of weekly volatility to annual risk. Our process is well modeled by fractional Gaussian noise. But this formula I do not understand. Thank you.
