[R] Problems with new arima function in ts

[kjetil halvorsen]

I have a model & data where arima0 works, but not the new arima:

> PIBconst
     Qtr1 Qtr2 Qtr3 Qtr4
1991 1599 1907 1927 1970
1992 1655 1894 1869 1952
1993 1711 1994 1892 1941
1994 1765 1936 2064 1969
1995 1834 1982 2005 2089
1996 1867 2059 2099 2075
1997 1872 2158 2114 2186
1998 1978 2175 2167 2226
1999 1971 2152 2117 2161
2000 1955 2167 2157 2206
2001 2025 2212 2175 2226

test <- arima(PIBconst, order=c(0,1,1), sea=list(order=c(0,1,1)),
init=c(0.5, 0.5))
Error in optim(init[mask], armaCSS, method = "BFGS", hessian = FALSE,  : 
        initial value in vmmin is not finite

with better init values:

 test <- arima(PIBconst, order=c(0,1,1), sea=list(order=c(0,1,1)),
init=c(-0.7, -0.8))
Error in optim(init[mask], armaCSS, method = "BFGS", hessian = FALSE,  : 
        initial value in vmmin is not finite

changing method:

 test <- arima(PIBconst, order=c(0,1,1), sea=list(order=c(0,1,1)),
init=c(-0.7, -0.8), method="ML")
Error in optim(init[mask], armafn, method = "BFGS", hessian = TRUE,
control = optim.control,  : 
        initial value in vmmin is not finite

What to do? (We have got this to work on other series, which actually
contain missing values).

Kjetil Halvorsen
-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-.-
r-help mailing list -- Read http://www.ci.tuwien.ac.at/~hornik/R/R-FAQ.html
Send "info", "help", or "[un]subscribe"
(in the "body", not the subject !)  To: r-help-request at stat.math.ethz.ch
_._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._._