I have to thank you for your several references but I must admit I could not understand exactly the argument...
My problem is about understanding deeply the differences between UK and IRF-k approaches...
It seems both of them see at the variable as the sum of a varying trend and a stationary residual; both of them consider such trend as the linear combination of monomials in points positions with some specific coefficients...suddenly all the references speak about the conditions for the coefficients of the IRF-k approach...that is their sum must be zero...the number of conditions depends on the degree of the polynomial trend...if such degree is zero, the only condition that satisfies the sum of coefficients equal to zero is that such coefficients are 1 and -1...that is the zero order increment or intrinsic approach (the stationarity of the increment)...now...
Neither one condition of this kind for UK?!...
Ho to pass from the trend to the conditions?...
Such trend is considered locally (within such neighbourhood) or globally? And in UK?...is it considered globally?
The universality condition in UK is exactly the same of the zero order one in Irf-k?...what are the differences?
How to introduce, in such framework, the order k intrinsic function?...what is it?...the combination of the original non stationary function with the said coefficients?...
Why such coefficients should filter higher order trends? In which way?
While in UK is intuitive to understand the concept of trend and residual (I try to fit the trend, I remove it from my original variable and I infer variogram on the residual), in Irf-k it is not!...
Which is the trend and which is the residual? Is the generalized covariance computed on such residual?
Dr. Simone Sammartino
- Geostatistical analyst
- G.I.S. mapping
I.A.M.C. - C.N.R.
Port of Naples - Naples
6X velocizzare la tua navigazione a 56k? 6X Web Accelerator di Libero!
Scaricalo su INTERNET GRATIS 6X http://www.libero.it