*
* Example 6.5.4 from pp 206-208
*
open data deaths.dat
calendar(m) 1973:1
data(format=free,org=columns) 1973:1 1978:12 deaths
*
* Use BJIDENT to get the autocorrelations and partial autocorrelations of the
* various combinations.
*
graph
# deaths
@bjident(diffs=1,sdiffs=1) deaths
*
* Estimate the model with first and seasonal differencing. To get the
* multiplicative structure with RATS, you just need to indicate one MA and one
* SMA (seasonal MA).
*
boxjenk(diff=1,sdiff=1,demean,ma=1,sma=1,maxl) deaths
@bdindtests %resids
*
* Estimate the full 13 lag MA model
*
boxjenk(diff=1,sdiff=1,demean,ma=13,iters=80) deaths
*
* Estimate a subset model
*
boxjenk(diff=1,sdiff=1,demean,ma=||1,6,12,13||,maxl) deaths