如何在一张图中绘制多个数据集的概率密度函数图？

Question

我为泊松分布生成了 100 个样本，样本大小为 n = 100、lambda = 3。现在我正在尝试在同一张图中绘制所有 100 个样本的概率密度函数图。

set.seed(124)


Sample100 <- replicate(100, matrix(data=rpois(100, lambda = 3),ncol = 1), simplify = TRUE)

View(Sample100)

colMeans(Sample100)
apply(Sample100, 2, var)

plot(density(Sample100),xlab = "y", ylab = "P(Y)",main = "Probability density function for 100 sample" )

我从代码中得到的是它只为我提供 1 行假设我将在图中得到 100 行，因为我有 100 组数据

Answer 1

如果你熟悉 ggplot 你可以这样做

set.seed(124)
library(ggplot2)
library(tidyr)
Sample100 <- replicate(100, matrix(data=rpois(100, lambda = 3),ncol = 1), simplify = TRUE)
Samplelong <- gather(as.data.frame(Sample100))

ggplot(Samplelong, aes(x = value)) +
       geom_density(aes(group = key)) +
       labs(x = "y", #labs function overides labels on the plot
            y = "P(y)",
            title = "Probability density function for 100 sample") +
       theme_classic() +   # This changes the theme to look similar to base r plot theme
       #theme function allows you to modify plot further, everyting from size
       # fonts, angles, axis.
       theme(plot.title = element_text(hjust = 0.5, #hjust - horizontal shift
                                       face = "bold")) + . # face - the print for plot.title is set to bold
       scale_x_continuous(expand = c(0,0))  + #for the continuous x axis, limit expansion to 0 on both ends
       scale_y_continuous(expand = c(0,0)) #for the continuous y axis, limit expansion to 0 on both ends

供参考，您似乎正在绘制以下内容

ggplot(Samplelong, aes(x = value)) +
       geom_density() +
       labs(x = "y", y = "P(y)", title = "Probability density function for 100 sample") +
       theme_classic() +
       theme(plot.title = element_text(hjust = 0.5)) +
       scale_x_continuous(expand = c(0,0))  +
       scale_y_continuous(expand = c(0,0))

如果 ggplot2 让您感到困惑，请花点时间尝试使图表更漂亮，对于大多数目的来说，以下是可以的。

 ggplot(Samplelong, aes(x = value)) +
   geom_density(aes(group = key)) +
   labs(x = "y", #labs function overides labels on the plot
        y = "P(y)",
        title = "Probability density function for 100 sample")