我有3个试验(试验：e1，e2，e3)，2个产品系列(产品：A，B)，5个率产品(.1,1,10,100,1000)，共6条曲线(曲线：c1,...,c6).在拟合一个非线性混合模型后，我想把所有的曲线和模型产生的曲线绘制在同一个图表中.这是参考资料(github中的包medrc): https:/doseresponse.github.iomedrcarticlesmedrc.html。
这是生成非线性混合模型的代码。
#packages
library(drc)
library(medrc)
library(dplyr)
library(tidyr)
#my data
trial <- c("e1","e1","e1","e1","e1","e1","e1","e1","e1","e1","e1","e1","e1","e1","e1",
"e1","e1","e1","e1","e1","e1","e1","e1","e1","e1","e1","e1","e1","e1","e1",
"e2","e2","e2","e2","e2","e2","e2","e2","e2","e2","e2","e2","e2","e2","e2",
"e2","e2","e2","e2","e2","e2","e2","e2","e2","e2","e2","e2","e2","e2","e2",
"e3","e3","e3","e3","e3","e3","e3","e3","e3","e3","e3","e3","e3","e3","e3",
"e3","e3","e3","e3","e3","e3","e3","e3","e3","e3","e3","e3","e3","e3","e3")
curve <- c("c1","c1","c1","c1","c1","c1","c1","c1","c1","c1","c1","c1","c1","c1","c1",
"c2","c2","c2","c2","c2","c2","c2","c2","c2","c2","c2","c2","c2","c2","c2",
"c3","c3","c3","c3","c3","c3","c3","c3","c3","c3","c3","c3","c3","c3","c3",
"c4","c4","c4","c4","c4","c4","c4","c4","c4","c4","c4","c4","c4","c4","c4",
"c5","c5","c5","c5","c5","c5","c5","c5","c5","c5","c5","c5","c5","c5","c5",
"c6","c6","c6","c6","c6","c6","c6","c6","c6","c6","c6","c6","c6","c6","c6")
rates <- c(.1,.1,.1,1,1,1,10,10,10,100,100,100,1000,1000,1000,
.1,.1,.1,1,1,1,10,10,10,100,100,100,1000,1000,1000,
.1,.1,.1,1,1,1,10,10,10,100,100,100,1000,1000,1000,
.1,.1,.1,1,1,1,10,10,10,100,100,100,1000,1000,1000,
.1,.1,.1,1,1,1,10,10,10,100,100,100,1000,1000,1000,
.1,.1,.1,1,1,1,10,10,10,100,100,100,1000,1000,1000)
product <- c("A","A","A","A","A","A","A","A","A","A","A","A","A","A","A",
"B","B","B","B","B","B","B","B","B","B","B","B","B","B","B",
"A","A","A","A","A","A","A","A","A","A","A","A","A","A","A",
"B","B","B","B","B","B","B","B","B","B","B","B","B","B","B",
"A","A","A","A","A","A","A","A","A","A","A","A","A","A","A",
"B","B","B","B","B","B","B","B","B","B","B","B","B","B","B")
resp <- c(.295,.3232,.3015,.155,.1501,.1483,.0511,.036,.0445,.0021,.0022,.0035,.0015,.0025,.0009,
.312,.3373,.2994,.265,.2501,.2482,.1022,.103,.1142,.0220,.0198,.0159,.0036,.0099,.0100,
.289,.3122,.3093,.141,.1612,.1398,.0722,.022,.0581,.0019,.0015,.0011,.0018,.0009,.0014,
.325,.3451,.2952,.267,.2412,.2398,.1125,.109,.1019,.0554,.0547,.0118,.0029,.0075,.0078,
.294,.2452,.2991,.121,.1925,.1485,.0871,.025,.0658,.0019,.0019,.0010,.0025,.0019,.0008,
.285,.3412,.3069,.124,.1861,.1958,.1276,.132,.1985,.0325,.0201,.0225,.0031,.0089,.0094)
data.test <- data.frame(trial,curve,rates,product,resp) #my data frame
#my model
m1 <- medrm(resp ~ rates,
curveid=b + c + d + e ~ product,
data = data.test,
fct=LL.4(),
random = c + d ~ 1|trial,
start=NULL)
为了使情节。
#plotting
pdata <- data.test%>%
group_by(curve, product) %>%
expand(rates=exp(seq(-3, 10, length=50)))
#pdata$resp_ind <- predict(m1, newdata=pdata)
pdata$resp <- predict(m1, newdata=pdata, level=0)
ggplot(data.test, aes(x=log(rates), y=resp,
colour=product, group=curve, shape=product)) +
geom_point() +
geom_line(data=pdata) +
#geom_line(data=pdata, aes(y=resp_ind), linetype=2) +
theme_bw() +
scale_x_continuous("DOSE",
breaks=log(c(.1, 1, 10, 100, 1000)),
labels=c(.1, 1, 10, 100, 1000))
注意，有两行代码是有注释的。在提取每条曲线的预测数据时，我无法指定水平，即给出随机分量的曲线。我错过了什么？
pdata$resp_ind <- predict(m1, newdata=pdata)
是导致错误。
Error in predict.nlme(object$fit, newdata = newdata, level = level) :
cannot evaluate groups for desired levels on 'newdata'
所以我不能用这一行代码来绘制每一条曲线线
geom_line(data=pdata, aes(y=resp_ind), linetype=2) +
这些都是类似的问题，但我用代码得到的是平均趋势。
pdata$resp <- predict(m1, newdata=pdata, level=0)
我想指定级别来获得所有的曲线。R: lme，无法评估'newdata'上所需级别的组。https:/stats.stackexchange.comquestions58031prediction-on-mixed-effect-models-what-to-do-with-random-effects。