将密度曲线拟合成R中的直方图

Question

将密度曲线拟合成R中的直方图

R中是否有与直方图拟合曲线的函数？

假设你有下面的直方图

hist(c(rep(65, times=5), rep(25, times=5), rep(35, times=10), rep(45, times=4)))

它看起来很正常，但它是倾斜的。我想要拟合一个正常的曲线，是倾斜的，围绕这个直方图。

这个问题是相当基本的，但我似乎无法在互联网上找到R的答案。

Answer 1

如果我正确理解了您的问题，那么您可能需要一个密度估计和直方图：X&nbsp;<-&nbsp;c(rep(65,&nbsp;times=5),&nbsp;rep(25,&nbsp;times=5),&nbsp;rep(35,&nbsp;times=10),&nbsp;rep(45,&nbsp;times=4))hist(X,&nbsp;prob=TRUE)&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; #&nbsp;prob=TRUE&nbsp;for&nbsp;probabilities&nbsp;not&nbsp;countslines(density(X))&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp; &nbsp;#&nbsp;add&nbsp;a&nbsp;density&nbsp;estimate&nbsp;with&nbsp;defaultslines(density(X,&nbsp;adjust=2),&nbsp;lty="dotted")&nbsp;&nbsp;&nbsp; &nbsp;#&nbsp;add&nbsp;another&nbsp;"smoother"&nbsp;density稍后编辑很长时间：下面是一个稍微打扮一下的版本：X&nbsp;<-&nbsp;c(rep(65,&nbsp;times=5),&nbsp;rep(25,&nbsp;times=5),&nbsp;rep(35,&nbsp;times=10),&nbsp;rep(45,&nbsp;times=4))hist(X,&nbsp;prob=TRUE,&nbsp;col="grey") #&nbsp;prob=TRUE&nbsp;for&nbsp;probabilities&nbsp;not&nbsp;countslines(density(X),&nbsp;col="blue",&nbsp;lwd=2) &nbsp;#&nbsp;add&nbsp;a&nbsp;density&nbsp;estimate&nbsp;with&nbsp;defaultslines(density(X,&nbsp;adjust=2),&nbsp;lty="dotted",&nbsp;col="darkgreen",&nbsp;lwd=2)

Answer 2

这样的事情用ggplot 2很容易。library(ggplot2)dataset&nbsp;<-&nbsp;data.frame(X&nbsp;=&nbsp;c(rep(65,&nbsp;times=5),&nbsp;rep(25,&nbsp;times=5),&nbsp; &nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;&nbsp;rep(35,&nbsp;times=10),&nbsp;rep(45,&nbsp;times=4)))ggplot(dataset,&nbsp;aes(x&nbsp;=&nbsp;X))&nbsp;+&nbsp; &nbsp;&nbsp;geom_histogram(aes(y&nbsp;=&nbsp;..density..))&nbsp;+&nbsp; &nbsp;&nbsp;geom_density()或者模仿Dirk的解的结果ggplot(dataset,&nbsp;aes(x&nbsp;=&nbsp;X))&nbsp;+&nbsp; &nbsp;&nbsp;geom_histogram(aes(y&nbsp;=&nbsp;..density..),&nbsp;binwidth&nbsp;=&nbsp;5)&nbsp;+&nbsp; &nbsp;&nbsp;geom_density()

Answer 3

我就是这样做的：foo&nbsp;<-&nbsp;rnorm(100,&nbsp;mean=1,&nbsp;sd=2)hist(foo,&nbsp;prob=TRUE)curve(dnorm(x,&nbsp;mean=mean(foo),&nbsp;sd=sd(foo)),&nbsp;add=TRUE)一个额外的练习就是用ggplot 2软件包来做这件事.。