Conclusions (4)
Evidence of “biscaling” in some traces:
- aɭ for mid-range scales
- aə for the coarsest scales
- it is due to distinct behavior of the PSD in various frequency bands
For the finest scales:
- correlations among three processes (I, P, B)
Length of the traces may be an issue
It may be that the data is non-stationary without deterministic trends.