Hi R users,
I'm here trying to understand correlated residuals in nonlinear estimation.
I'm reading/studying the book Bates, D. M. and D. G. Watts, (1988),
/Nonlinear regression analysis and its applications/, Wiley, NY. pages
92-94, trying to reproduce the figures and to find out the code in
library(nlme)
m2 - gnls(conc ~ t1*(1-t2*exp(-k*time)),
data = df.Chloride,
start = list(
t1 = 35,
t2 = 0.91,
k = 0.22))
summary(m2)
plot(m2)
lag.plot(resid(m2), do.lines=FALSE)
acf(resid(m2))
m3 - update(m2, corr=corAR1(.67))
2 matches
Mail list logo