Is there a library dealing with correlation in the residuals of a glm?
I have
bin3alt <-glm(respalt~ t+sn+c5.vrm,data=dfalt,family="quasibinomial")
> bin3alt
Call: glm(formula = respalt ~ t + sn + c5.vrm, family = "quasibinomial",
data = dfalt)
Coefficients:
(Intercept) t2 t3 t4 t5 t6
sn2 sn3 c5.vrm
-3.35957 1.81455 0.96161 -0.37701 -2.32657 -3.75074
0.24266 0.39056 0.06673
Degrees of Freedom: 230 Total (i.e. Null); 222 Residual
Null Deviance: 107000
Residual Deviance: 2290 AIC: NA
dfalt$pears <- residuals(bin3alt,type="pearson")
arima(dfalt$pears[dfalt$t==4],order=c(1,0,0)))
Call:
arima(x = dfalt$pears[dfalt$t == 4], order = c(1, 0, 0))
Coefficients:
ar1 intercept
0.6333 -0.5091
s.e. 0.1257 1.0490
Not all levels of the t factor, show correlation, but some do. The factor is
not a random effect it is month of ageing. Also, if I use the Cochrane Orcutt
manually, should I use response or pearson residuals? I know of lme, but think
it requires a random effect.
Stephen Bond
[[alternative HTML version deleted]]
______________________________________________
R-help@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-help
PLEASE do read the posting guide http://www.R-project.org/posting-guide.html
and provide commented, minimal, self-contained, reproducible code.
