如何为每个人添加趋势线，以便在Base R中不使用循环？

我已经改编了一些非常漂亮的代码（参见下面 - 感谢@Ben-Bolker）使用mapply创建一个绘图（请参见下文），以便我不必使用循环。我想用类似的技术为每个人添加一条趋势线到现有的地块。有什么建议么？如何为每个人添加趋势线，以便在Base R中不使用循环？

enter image description here

下面的图（下面的代码）类似：

enter image description here

## sample data ## 
WW_Wing_SI <- structure(list(Individual = c("WW_08A_08", "WW_08A_08", "WW_08A_08", 
"WW_08A_08", "WW_08A_08", "WW_08A_08", "WW_08A_08", "WW_08A_08", 
"WW_08A_08", "WW_08B_02", "WW_08B_02", "WW_08B_02", "WW_08B_02", 
"WW_08B_02", "WW_08B_02", "WW_08B_02", "WW_08B_02", "WW_08B_02", 
"WW_08G_01", "WW_08G_01", "WW_08G_01", "WW_08G_01", "WW_08G_01", 
"WW_08G_01", "WW_08G_01", "WW_08G_01", "WW_08G_01", "WW_08G_05", 
"WW_08G_05", "WW_08G_05", "WW_08G_05", "WW_08G_05", "WW_08G_05", 
"WW_08G_05", "WW_08G_05", "WW_08G_05"), Feather = c("1", "2", 
"3", "4", "5", "6", "7", "8", "9", "1", "2", "3", "4", "5", "6", 
"7", "8", "9", "1", "2", "3", "4", "5", "6", "7", "8", "9", "1", 
"2", "3", "4", "5", "6", "7", "8", "9"), Delta15N = c(8.26, 8.1, 
8.07, 8.7, 8.98, 9.44, 7.84, 7.26, 6.05, 6.9, 6.73, 6.97, 6.67, 
6.76, 6.59, 6.58, 6.42, 6.3, 11.64, 11.83, 11.66, 11.3, 11.32, 
11.29, 10.91, 10.77, 11.4, 7.7, 7.8, 8.29, 9.65, 10.25, 13.67, 
14.66, 13.48, 13.76)), .Names = c("Individual", "Feather", "Delta15N" 
), row.names = c(NA, 36L), class = "data.frame") 

## plot delta15N by feather position for each individual ## 
xvals <- tapply(WW_Wing_SI$Feather, WW_Wing_SI$Individual,function(x) return(x)) 
yvals <- tapply(WW_Wing_SI$Delta15N,WW_Wing_SI$Individual, function(x) return(x)) 
ID <- unique(WW_Wing_SI$Individual) 

par(oma = c(2, 2, 0, 0), mar = c(4, 5, 2, 2), pty = "s") 
plot(1:max(unlist(xvals)),ylim=(c(floor(min(unlist(yvals))),ceiling(max(unlist(yvals))))),type="n", bty = "n", 
    cex.lab = 1.75, cex.axis = 1.75, main = NULL, axes = F, 
    xlab="Feather Position", ylab=expression(paste(delta ^{15},"N"))) 
axis(1, at = seq(1, 9, by = 1), labels = T, tick = TRUE, cex.axis = 1.25, cex.lab = 1.25, lwd = 1.25, lwd.ticks = 1.25) 
axis(2, at = seq(floor(min(unlist(yvals))), ceiling(max(unlist(yvals))), by = 1), labels = T, tick = TRUE, cex.axis = 1.25, cex.lab = 1.25, lwd = 1.25, lwd.ticks = 1.25) 
mapply(lines,xvals, yvals, col = c(1:nrow(xvals)), pch = c(1:nrow(xvals)), type = "o", lty = c(1:nrow(xvals))) 

legend("bottom", ID, pch = c(1:nrow(xvals)), col = c(1:nrow(xvals)), cex = 1, pt.bg = c(1:nrow(xvals)), horiz=TRUE, bty = "n") 


## Code to accomplish using ggplot ## 
WW_Wing_SI$Feather <- as.numeric(WW_Wing_SI$Feather) 
library(ggplot2) 
theme_set(theme_bw()) 
ggplot(WW_Wing_SI,aes(Feather,Delta15N,fill=Individual,colour=Individual))+ 
    geom_line()+geom_smooth(method="lm",formula=y~poly(x,2),linetype=2)

来源

2013-02-08 Keith W. Larson

你想避免只是为了美观代码原因循环？据我所知，'mapply'（和'apply'函数的整个系列）是一个循环的包装函数。 – Dinre 2013-02-08 15:51:58

简单的代码？我只是假设必须有类似的方式来添加趋势线，就像我通过线连接的点一样。 – 2013-02-08 15:53:27

'matplot'怎么样？（这里有创意:-)）。我经常用第一列的x值和另一列的每个变量构建一个数组。 – 2013-02-08 16:04:15

这里有一个建议。首先将线性模型拟合到您的数据中（使用lm）。然后使用这些拟合绘制线：

fits <- mapply(function(x, y) lm(y ~ as.numeric(x)), 
       xvals, yvals, SIMPLIFY = FALSE) 
mapply(abline, fits, col = seq_along(xvals))

enter image description here

来源

2013-02-08 16:05:15

在优秀的代码中是否可以使用平滑的线条？ – 2013-02-08 16:06:59

@KeithLarson你是什么意思的“平滑线”？非线性拟合？ – 2013-02-08 16:07:58

是的，非线性的，例如WW_08A_08。 – 2013-02-08 16:09:01

如何为每个人添加趋势线，以便在Base R中不使用循环？

回答

相关问题