3

我试图在 ggplot2 中绘制来自具有相同预测变量的两个线性模型的回归变量系数的多面图。我构建的数据框是这样的： 

r.together>
          reg         coef        se      y
1  (Intercept)  5.068608671 0.6990873 Labels
2     goodTRUE  0.310575129 0.5228815 Labels
3    indiaTRUE -1.196868662 0.5192330 Labels
4    moneyTRUE -0.586451273 0.6011257 Labels
5     maleTRUE -0.157618168 0.5332040 Labels
6  (Intercept)  4.225580743 0.6010509  Bonus
7     goodTRUE  1.272760149 0.4524954  Bonus
8    indiaTRUE -0.829588862 0.4492838  Bonus
9    moneyTRUE -0.003571476 0.5175601  Bonus
10    maleTRUE  0.977011737 0.4602726  Bonus

“y”列是模型的标签，reg 是回归量，coef 和 se 是您的想法。

我想绘制：

g <- qplot(reg, coef, facets=.~y, data = r.together) + coord_flip()

但是当我尝试显示情节时，我得到：

> print(g)
Error in names(df) <- output : 
'names' attribute [2] must be the same length as the vector [1]

奇怪的是

qplot(reg, coef, colour=y, data = r.together) + coord_flip()

情节如你所料。

4

2 回答 2

5

我不知道为什么，但名字y似乎是个问题。如果你改变它，那么它的工作原理

r.together <- read.table(textConnection("
          reg         coef        se      myfactor
 (Intercept)  5.068608671 0.6990873 Labels
    goodTRUE  0.310575129 0.5228815 Labels
   indiaTRUE -1.196868662 0.5192330 Labels
   moneyTRUE -0.586451273 0.6011257 Labels
    maleTRUE -0.157618168 0.5332040 Labels
 (Intercept)  4.225580743 0.6010509  Bonus
    goodTRUE  1.272760149 0.4524954  Bonus
   indiaTRUE -0.829588862 0.4492838  Bonus
   moneyTRUE -0.003571476 0.5175601  Bonus
    maleTRUE  0.977011737 0.4602726  Bonus
"),header=T)

qplot(reg, coef, data = r.together, facets= .~myfactor ) + coord_flip()
于 2010-04-11T17:49:14.260 回答
2

可以是保留字吗？

r.together=data.frame(
  reg=c("(Intercept)", "goodTRUE", "indiaTRUE", "moneyTRUE", "maleTRUE", "(Intercept)", "goodTRUE", "indiaTRUE", "moneyTRUE", "maleTRUE"),
  coef=c(5.068608671, 0.310575129, -1.196868662, -0.586451273, -0.157618168, 4.225580743, 1.272760149, -0.829588862, -0.003571476, 0.977011737),
  se=c(0.6990873, 0.5228815, 0.519233, 0.6011257, 0.533204, 0.6010509, 0.4524954, 0.4492838, 0.5175601, 0.4602726),
  yy= c("Labels", "Labels", "Labels", "Labels", "Labels", "Bonus", "Bonus", "Bonus", "Bonus", "Bonus")
 )

yy 似乎有效

g <- qplot(reg, coef, data = r.together,facets=. ~ yy) +coord_flip()
于 2010-04-11T17:58:09.093 回答