r - 适用于任意数量的结/断点的可扩展分段函数

Question

我有以下斜率、转折点和截距的集合:

slopes <- c(4, 2, 8, 4)
breaks <- c(0.0150, 0.030, 0.035)
intercepts <- c(0.0299, 0.0599, -0.1201, 0.0199)

它们定义了以下几行:

# y = slopes[1] * x + intercepts[1]
# y = slopes[2] * x + intercepts[2]
# y = slopes[3] * x + intercepts[3]
# y = slopes[4] * x + intercepts[4]

绘制线条产量:

tibble(x = seq(0.0025, 0.06, 0.0025), y = x) %>% 
    ggplot(aes(x, y)) +
    geom_point(alpha = 0) + 
    geom_abline(intercept = intercepts[1], slope = slopes[1], color = "red") + 
    geom_abline(intercept = intercepts[2], slope = slopes[2], color = "orange") + 
    geom_abline(intercept = intercepts[3], slope = slopes[3], color = "yellow") +
    geom_abline(intercept = intercepts[4], slope = slopes[4], color = "green2") +
    scale_y_continuous(limits = c(0, 1))

我想创建一个基于线和断点/结点的分段函数，如下所示(遵循:红色 -> 橙色 -> 黄色 -> 绿色):

我可以将一个函数包装在几个 if/else 语句上以获得我想要的。但我希望该解决方案可以针对任意数量的断裂/结进行扩展(在此示例中，而不是仅仅 3 个)。

我怎样才能做到这一点？

Answer 1

这应该是相当可扩展的:

piecewise <- function(x, slopes, intercepts, breaks) {
    i = 1 + findInterval(x, breaks)
    y = slopes[i] * x + intercepts[i]
    return(y)
}

请注意，我将 breaks 参数放在最后，因为这对我来说似乎最自然。它自动为任意数量的片段实现分段定义的函数。

例子:

slopes <- c(4, 2, 8, 4)
intercepts <- c(0.0299, 0.0599, -0.1201, 0.0199)
breaks <- c(0.0150, 0.030, 0.035)

df <- tibble(x = seq(0.0025, 0.06, 0.0025)) %>% 
    mutate(y = piecewise(x, slopes, intercepts, breaks))

df %>% 
    ggplot(aes(x, y)) + 
    geom_line()