将线段添加到密度图答案

【问题标题】：Add line segment to density plot将线段添加到密度图
【发布时间】：2022-06-11 02:43:55
【问题描述】：

我想在我的图中添加一条线来表示平均值：

X1 <- rnorm(100)
# Kernel density estimates
density_X1 <- density(X1)
# Compute mode
mode_X1 <- density_X1$x[which.max(density_X1$y)]; mode_X1
# Compute mean
mean_X1 <- mean(X1)
# Create plot 
plot(density_X1, main = "Kernel density estimates", xlab = "X1", ylab = "density")
# Add line for mode
segments(x0 = mode_X1, y0 = 0, x1 = mode_X1, y1 = density_X1$y, col = "red", lty = 1, lwd = 1)
# Add line for mean
segments(x0 = mean_X1, y0 = 0, x1 = mean_X1, y1 = , col = "red", lty = 5, lwd = 2)

我需要为segments(y1 = ) 输入什么值以确保上限是核密度图？

【问题讨论】：

标签： r plot

【解决方案1】：

要根据离散点的密度估计值计算某个 x 值的 y 值，您需要进行某种插值。一种方法是使用 approx 的线性插值：

approx(x = density_X1$x, y = density_X1$y, xout = mean_X1)$y

其中，x 和 y 给出要插值的点，xout 是完成插值的 x 值的（向量）（参见 ?approx）

代码如下：

set.seed(72405277)
X1 <- rnorm(100)
# Kernel density estimates
density_X1 <- density(X1)
# Compute mode
mdIdx <- which.max(density_X1$y) # store and resuse
mode_X1 <- density_X1$x[mdIdx]; 
# Compute mean
mean_X1 <- mean(X1)
# Create plot 
plot(density_X1, main = "Kernel density estimates", xlab = "X1", ylab = "density")
# Add line for mode
segments(x0 = mode_X1, y0 = 0, x1 = mode_X1, y1 = density_X1$y[mdIdx], col = "red", lty = 1, lwd = 1)
# Add line for mean
yax2 <- approx(x = density_X1$x, y = density_X1$y, xout = mean_X1)$y
segments(x0 = mean_X1, y0 = 0, x1 = mean_X1, y1 = yax2, col = "blue", lty = 5, lwd = 2)

【讨论】：

你也可以试试kde1d 包。然后，该软件包具有可以轻松估计特定值（以及更多设施）的密度的功能。所以你可以通过d <- kde1d(X1); dkde1d(mean(X1), d) 来获取值（注意，由于密度估计器不同，这与上面有点不同）

【解决方案2】：

你也可以像这样使用abline：

X1 <- rnorm(100)
# Kernel density estimates
density_X1 <- density(X1)
# Compute mode
mode_X1 <- density_X1$x[which.max(density_X1$y)]; mode_X1
# Compute mean
mean_X1 <- mean(X1)
# Create plot 
plot(density_X1, main = "Kernel density estimates", xlab = "X1", ylab = "density")
abline(v = mode_X1, col = "red", lty = 1, lwd = 1)
abline(v = mean_X1, col = "red", lty = 5, lwd = 2)

输出：

【讨论】：