Grocer Function
NAME
transdif - differenciation and Box-Cox transformation
CALLING SEQUENCE
[yt]=transdif(y,lambda,d,ds,s)
PARAMETERS
Input
-
y = (nxk) data matrix
- lambda = a scalar, parameter of the Box-Cox transformation
- d = a scalar, the number of regular differences (1-B)^d
-
ds = (Sx1) matrix containing the number of seasonal differences
-
s = (rx1) matrix containing the seasonal periods
Output
- yt =(n-d-sum(ds*s))xk matrix of transformed data
DESCRIPTION
Function that mimics matlab function transdif: forms the string matrix
EXAMPLE
1. mtlb_load('SCI/scied/grocer/encours/e4sci/seriesa.dat'); elec_cons = transdif(seriesa,0,1,1,12);
2. mtlb_load('SCI/scied/grocer/encours/e4sci/seriesa.dat'); s = transdif(seriesa,-1,0,0,12);
Example 1 taken from varma function varma_d1(). Elec_cons is equal to (1-L)(1-L^12)log(seriesa).
Example 2 gives simply the box-cox transformation of seriesa with lambda = -1 (that is 1-1/seriesa)
AUTHOR
Jaime Terceiro, 1997/ Eric Dubois 2004