#-*- R -*- # initialization library(nlme) library(lattice) options(width = 65, ## reduce platform dependence in printed output when testing digits = if(nzchar(Sys.getenv("R_TESTS"))) 3 else 5) options(contrasts = c(unordered = "contr.helmert", ordered = "contr.poly")) pdf(file = 'ch04.pdf') # Chapter 4 Fitting Linear Mixed-Effects Models # 4.1 Fitting Linear Models in S with lm and lmList fm1Orth.lm <- lm(distance ~ age, Orthodont) fm1Orth.lm par(mfrow=c(2,2)) plot(fm1Orth.lm) # Figure 4.1 fm2Orth.lm <- update(fm1Orth.lm, formula = distance ~ Sex*age) summary(fm2Orth.lm) fm3Orth.lm <- update(fm2Orth.lm, formula = . ~ . - Sex) summary(fm3Orth.lm) bwplot(getGroups(Orthodont)~resid(fm2Orth.lm)) # Figure 4.2 fm1Orth.lis <- lmList(distance ~ age | Subject, Orthodont) getGroupsFormula(Orthodont) fm1Orth.lis <- lmList(distance ~ age, Orthodont) formula(Orthodont) fm1Orth.lis <- lmList(Orthodont) fm1Orth.lis summary(fm1Orth.lis) pairs(fm1Orth.lis, id = 0.01, adj = -0.5) # Figure 4.3 fm2Orth.lis <- update(fm1Orth.lis, distance ~ I(age-11)) intervals(fm2Orth.lis) plot(intervals(fm2Orth.lis)) # Figure 4.5 IGF plot(IGF) # Figure 4.6 fm1IGF.lis <- lmList(IGF) coef(fm1IGF.lis) plot(intervals(fm1IGF.lis)) # Figure 4.7 fm1IGF.lm <- lm(conc ~ age, data = IGF) summary(fm1IGF.lm) # 4.2 Fitting Linear Mixed-Effects Models with lme fm1Orth.lme <- lme(distance ~ I(age-11), data = Orthodont, random = ~ I(age-11) | Subject) fm1Orth.lme <- lme(distance ~ I(age-11), data = Orthodont) fm1Orth.lme <- lme(fm2Orth.lis) fm1Orth.lme fm2Orth.lme <- update(fm1Orth.lme, distance~Sex*I(age-11)) summary(fm2Orth.lme) fitted(fm2Orth.lme, level = 0:1) resid(fm2Orth.lme, level = 1) resid(fm2Orth.lme, level = 1, type = "pearson") newOrth <- data.frame(Subject = rep(c("M11","F03"), c(3, 3)), Sex = rep(c("Male", "Female"), c(3, 3)), age = rep(16:18, 2)) predict(fm2Orth.lme, newdata = newOrth) predict(fm2Orth.lme, newdata = newOrth, level = 0:1) fm2Orth.lmeM <- update(fm2Orth.lme, method = "ML") summary(fm2Orth.lmeM) compOrth <- compareFits(coef(fm2Orth.lis), coef(fm1Orth.lme)) compOrth plot(compOrth, mark = fixef(fm1Orth.lme)) # Figure 4.8 ## Figure 4.9 plot(comparePred(fm2Orth.lis, fm1Orth.lme, length.out = 2), layout = c(8,4), between = list(y = c(0, 0.5, 0))) plot(compareFits(ranef(fm2Orth.lme), ranef(fm2Orth.lmeM)), mark = c(0, 0)) fm4Orth.lm <- lm(distance ~ Sex * I(age-11), Orthodont) summary(fm4Orth.lm) anova(fm2Orth.lme, fm4Orth.lm) #fm1IGF.lme <- lme(fm1IGF.lis) #fm1IGF.lme #intervals(fm1IGF.lme) #summary(fm1IGF.lme) pd1 <- pdDiag(~ age) pd1 formula(pd1) #fm2IGF.lme <- update(fm1IGF.lme, random = pdDiag(~age)) (fm2IGF.lme <- lme(conc ~ age, IGF, random = pdDiag(~age))) #anova(fm1IGF.lme, fm2IGF.lme) anova(fm2IGF.lme) #update(fm1IGF.lme, random = list(Lot = pdDiag(~ age))) pd2 <- pdDiag(value = diag(2), form = ~ age) pd2 formula(pd2) lme(conc ~ age, IGF, pdDiag(diag(2), ~age)) fm4OatsB <- lme(yield ~ nitro, data = Oats, random =list(Block = pdCompSymm(~ Variety - 1))) summary(fm4OatsB) corMatrix(fm4OatsB$modelStruct$reStruct$Block)[1,2] fm4OatsC <- lme(yield ~ nitro, data = Oats, random=list(Block=pdBlocked(list(pdIdent(~ 1), pdIdent(~ Variety-1))))) summary(fm4OatsC) ## establishing the desired parameterization for contrasts options(contrasts = c("contr.treatment", "contr.poly")) fm1Assay <- lme(logDens ~ sample * dilut, Assay, random = pdBlocked(list(pdIdent(~ 1), pdIdent(~ sample - 1), pdIdent(~ dilut - 1)))) fm1Assay anova(fm1Assay) formula(Oxide) fm1Oxide <- lme(Thickness ~ 1, Oxide) fm1Oxide intervals(fm1Oxide, which = "var-cov") fm2Oxide <- update(fm1Oxide, random = ~ 1 | Lot) anova(fm1Oxide, fm2Oxide) coef(fm1Oxide, level = 1) coef(fm1Oxide, level = 2) ranef(fm1Oxide, level = 1:2) fm1Wafer <- lme(current ~ voltage + I(voltage^2), data = Wafer, random = list(Wafer = pdDiag(~voltage + I(voltage^2)), Site = pdDiag(~voltage + I(voltage^2)))) ## IGNORE_RDIFF_BEGIN summary(fm1Wafer) ## IGNORE_RDIFF_END fitted(fm1Wafer, level = 0) resid(fm1Wafer, level = 1:2) newWafer <- data.frame(Wafer = rep(1, 4), voltage = c(1, 1.5, 3, 3.5)) predict(fm1Wafer, newWafer, level = 0:1) newWafer2 <- data.frame(Wafer = rep(1, 4), Site = rep(3, 4), voltage = c(1, 1.5, 3, 3.5)) predict(fm1Wafer, newWafer2, level = 0:2) # 4.3 Examining a Fitted Model plot(fm2Orth.lme, Subject~resid(.), abline = 0) plot(fm2Orth.lme, resid(., type = "p") ~ fitted(.) | Sex, id = 0.05, adj = -0.3) fm3Orth.lme <- update(fm2Orth.lme, weights = varIdent(form = ~ 1 | Sex)) fm3Orth.lme plot(fm3Orth.lme, distance ~ fitted(.), id = 0.05, adj = -0.3) anova(fm2Orth.lme, fm3Orth.lme) qqnorm(fm3Orth.lme, ~resid(.) | Sex) plot(fm2IGF.lme, resid(., type = "p") ~ fitted(.) | Lot, layout = c(5,2)) qqnorm(fm2IGF.lme, ~ resid(.), id = 0.05, adj = -0.75) plot(fm1Oxide) qqnorm(fm1Oxide) plot(fm1Wafer, resid(.) ~ voltage | Wafer) plot(fm1Wafer, resid(.) ~ voltage | Wafer, panel = function(x, y, ...) { panel.grid() panel.xyplot(x, y) panel.loess(x, y, lty = 2) panel.abline(0, 0) }) with(Wafer, coef(lm(resid(fm1Wafer) ~ cos(4.19*voltage)+sin(4.19*voltage)-1))) nls(resid(fm1Wafer) ~ b3*cos(w*voltage) + b4*sin(w*voltage), Wafer, start = list(b3 = -0.0519, b4 = 0.1304, w = 4.19)) fm2Wafer <- update(fm1Wafer, . ~ . + cos(4.5679*voltage) + sin(4.5679*voltage), random = list(Wafer=pdDiag(~voltage+I(voltage^2)), Site=pdDiag(~voltage+I(voltage^2)))) summary(fm2Wafer) ## IGNORE_RDIFF_BEGIN intervals(fm2Wafer) ## IGNORE_RDIFF_END qqnorm(fm2Wafer) qqnorm(fm2Orth.lme, ~ranef(.), id = 0.10, cex = 0.7) pairs(fm2Orth.lme, ~ranef(.) | Sex, id = ~ Subject == "M13", adj = -0.3) fm2IGF.lme c(0.00031074, 0.0053722)/abs(fixef(fm2IGF.lme)) fm3IGF.lme <- update(fm2IGF.lme, random = ~ age - 1) anova(fm2IGF.lme, fm3IGF.lme) qqnorm(fm1Oxide, ~ranef(., level = 1), id=0.10) qqnorm(fm1Oxide, ~ranef(., level = 2), id=0.10) #fm3Wafer <- update(fm2Wafer, # random = list(Wafer = ~voltage+I(voltage^2), # Site = pdDiag(~voltage+I(voltage^2))), # control = list(msVerbose = TRUE, msMaxIter = 200) # ) #fm3Wafer #anova(fm2Wafer, fm3Wafer) #fm4Wafer <- update(fm2Wafer, # random = list(Wafer = ~ voltage + I(voltage^2), # Site = pdBlocked(list(~1, # ~voltage+I(voltage^2) - 1))), # control = list(msVerbose = TRUE, # msMaxIter = 200)) #fm4Wafer #anova(fm3Wafer, fm4Wafer) #qqnorm(fm4Wafer, ~ranef(., level = 2), id = 0.05, # cex = 0.7, layout = c(3, 1)) # The next line is not in the book but is needed to get fm1Machine fm1Machine <- lme(score ~ Machine, data = Machines, random = ~ 1 | Worker) (fm3Machine <- update(fm1Machine, random = ~Machine-1|Worker)) # cleanup summary(warnings())