用于广义帕累托分布的具有多条密度曲线的叠加直方图

Posted 2023-02-16

【中文标题】用于广义帕累托分布的具有多条密度曲线的叠加直方图

【英文标题】：Overlay histogram with multiple density curves for generalized pareto distribution

【发布时间】：2021-08-21 02:59:48

【问题描述】：

我想为广义帕累托分布绘制具有多条密度曲线的叠加直方图。如您所见，密度曲线不清晰可见。他们有什么办法说清楚吗？谢谢。

library(eva)
data(lowestoft)
data = as.vector(lowestoft)

d1 = dgpd(data, 0,  31.38105, 10.15003)
d2 = dgpd(data, 0, 2.9431553, 0.6778055)
d3 = dgpd(data,0, 5.413916, 17.162103)
d4 = dgpd(data,0, 43.18705, 13.98005)


N = length(data)
allden = c(d1, d2, d3, d4)
settings = c(rep('d1', N), rep('d2', N), 
             rep('d3', N), rep('d4', N))


mydata = data.frame(x= rep(data, 4), allden = allden, Methods = settings)

ggplot(mydata, aes(x)) +
  geom_histogram(aes(y = stat(density)), binwidth = 1, fill = "grey", color = "black")+
  geom_line(aes(x = x, y=allden, color = Methods)) 

【参考方案1】：

您可以做的一件事是像这样为geom_line 添加一个单独的缩放 y 轴

library(eva)
data(lowestoft)
data = as.vector(lowestoft)

d1 = dgpd(data, 0,  31.38105, 10.15003)
d2 = dgpd(data, 0, 2.9431553, 0.6778055)
d3 = dgpd(data,0, 5.413916, 17.162103)
d4 = dgpd(data,0, 43.18705, 13.98005)


N = length(data)
allden = c(d1, d2, d3, d4)
settings = c(rep('d1', N), rep('d2', N), 
             rep('d3', N), rep('d4', N))


mydata = data.frame(x= rep(data, 4), allden = allden, Methods = settings)

coef <- 8
ggplot(mydata, aes(x,allden)) +
  geom_histogram(aes(y = stat(density)), binwidth = 1, fill = "grey", color = "black") +
  geom_line(aes(x = x, y=allden*coef, color = Methods)) +
  scale_y_continuous(name="Histogram Axis",
                     sec.axis = sec_axis(trans~./coef, name='Line Axis'))