如何在dotplot或ggplot2中按随机效应的值（不是截距）对随机效果进行排序 [英] How can I sort random effects by value of the random effect (not the intercept) in dotplot or ggplot2

问题描述

我怀疑这个问题的答案很简单，我只是不知道它是什么。

$ b 长话短说，我想要显示一个点图我估计的模型中的随机截距和斜率。我使用 ggCaterpillar 函数有用地介绍了

例如，如果您愿意，可以在绘图之前对行进行重新排序以反转订单：

  ref < -  ranef（fit，condVar = TRUE）
 ref $ Subject< -  ref $ Subject [nrow（ref $ Subject）：1，] 
 ggCaterpillar（ref，QQ = FALSE，likeDotplot = TRUE，reorder = FALSE）[[Subject]] 
  

< img src =https://i.stack.imgur.com/7iTML.pngalt =在这里输入图片描述>

I suspect the answer to this question is rather simple and I just don't know what it is.

Long story short, I want to display a dotplot of the random intercepts and slopes from a model I am estimating. I'm using the ggCaterpillar function helpfully introduced here. However, this function, as well as the standard dotplot from lattice, sort the ensuing graph by decreasing order of the random intercept. I would like to sort the graph by increasing value of the random effect (either alphabetical or numerical).

Consider this minimum working example that comes standard in the lme4 package, along with the ggCaterpillar function.

## https://stackoverflow.com/questions/13847936/in-r-plotting-random-effects-from-lmer-lme4-package-using-qqmath-or-dotplot
ggCaterpillar <- function(re, QQ=TRUE, likeDotplot=TRUE) {
    require(ggplot2)
f <- function(x) {
    pv   <- attr(x, "postVar")
    cols <- 1:(dim(pv)[1])
    se   <- unlist(lapply(cols, function(i) sqrt(pv[i, i, ])))
    ord  <- unlist(lapply(x, order)) + rep((0:(ncol(x) - 1)) * nrow(x), each=nrow(x))
    pDf  <- data.frame(y=unlist(x)[ord],
                       ci=1.96*se[ord],
                       nQQ=rep(qnorm(ppoints(nrow(x))), ncol(x)),
                       ID=factor(rep(rownames(x), ncol(x))[ord], levels=rownames(x)[ord]),
                       ind=gl(ncol(x), nrow(x), labels=names(x)))

    if(QQ) {  ## normal QQ-plot
        p <- ggplot(pDf, aes(nQQ, y))
        p <- p + facet_wrap(~ ind, scales="free")
        p <- p + xlab("Standard normal quantiles") + ylab("Random effect quantiles")
    } else {  ## caterpillar dotplot
        p <- ggplot(pDf, aes(ID, y)) + coord_flip()
        if(likeDotplot) {  ## imitate dotplot() -> same scales for random effects
            p <- p + facet_wrap(~ ind)
        } else {           ## different scales for random effects
            p <- p + facet_grid(ind ~ ., scales="free_y")
        }
        p <- p + xlab("Levels") + ylab("Random effects")
    }

    p <- p + theme(legend.position="none")
    p <- p + geom_hline(yintercept=0)
    p <- p + geom_errorbar(aes(ymin=y-ci, ymax=y+ci), width=0, colour="black")
    p <- p + geom_point(aes(size=1.2), colour="blue") 
    return(p)
}

lapply(re, f)
}

library(lme4)
fit <- lmer(Reaction ~ Days + (Days|Subject), sleepstudy)
ggCaterpillar(ranef(fit,condVar=TRUE), QQ=FALSE, likeDotplot=TRUE)[["Subject"]] 

The graph I get looks like this.

How can I order the graph so the plots are ordered by increasing value of the random effect (e.g. 308, 309, 310... in the sleepstudy case)?

解决方案

In the part of the function where ord and pDf are created the rows are reordered by size of the intercept. We can make that part optional like in the updated function below:

require(ggplot2)
ggCaterpillar <- function(re, QQ=TRUE, likeDotplot=TRUE, reorder=TRUE) {
  require(ggplot2)
  f <- function(x) {
    pv   <- attr(x, "postVar")
    cols <- 1:(dim(pv)[1])
    se   <- unlist(lapply(cols, function(i) sqrt(pv[i, i, ])))
    if (reorder) {
      ord  <- unlist(lapply(x, order)) + rep((0:(ncol(x) - 1)) * nrow(x), each=nrow(x))
      pDf  <- data.frame(y=unlist(x)[ord],
                         ci=1.96*se[ord],
                         nQQ=rep(qnorm(ppoints(nrow(x))), ncol(x)),
                         ID=factor(rep(rownames(x), ncol(x))[ord], levels=rownames(x)[ord]),
                         ind=gl(ncol(x), nrow(x), labels=names(x)))
    } else {
      pDf  <- data.frame(y=unlist(x),
                         ci=1.96*se,
                         nQQ=rep(qnorm(ppoints(nrow(x))), ncol(x)),
                         ID=factor(rep(rownames(x), ncol(x)), levels=rownames(x)),
                         ind=gl(ncol(x), nrow(x), labels=names(x)))
    }

    if(QQ) {  ## normal QQ-plot
      p <- ggplot(pDf, aes(nQQ, y))
      p <- p + facet_wrap(~ ind, scales="free")
      p <- p + xlab("Standard normal quantiles") + ylab("Random effect quantiles")
    } else {  ## caterpillar dotplot
      p <- ggplot(pDf, aes(ID, y)) + coord_flip()
      if(likeDotplot) {  ## imitate dotplot() -> same scales for random effects
        p <- p + facet_wrap(~ ind)
      } else {           ## different scales for random effects
        p <- p + facet_grid(ind ~ ., scales="free_y")
      }
      p <- p + xlab("Levels") + ylab("Random effects")
    }

    p <- p + theme(legend.position="none")
    p <- p + geom_hline(yintercept=0)
    p <- p + geom_errorbar(aes(ymin=y-ci, ymax=y+ci), width=0, colour="black")
    p <- p + geom_point(aes(size=1.2), colour="blue") 
    return(p)
  }

  lapply(re, f)
}

If we now call the function with reorder = FALSE we get the original order of the rows:

ggCaterpillar(ranef(fit,condVar=TRUE), QQ=FALSE, likeDotplot=TRUE, reorder=FALSE)[["Subject"]]

You can reorder rows before plotting if you'd like, for example to reverse the order:

ref <- ranef(fit,condVar=TRUE)
ref$Subject <- ref$Subject[nrow(ref$Subject):1, ]
ggCaterpillar(ref, QQ=FALSE, likeDotplot=TRUE, reorder=FALSE)[["Subject"]]

这篇关于如何在dotplot或ggplot2中按随机效应的值（不是截距）对随机效果进行排序的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！