非线性拟合 ggplot2 R 中的 facet_grid 函数

Question

对于 X 列，所有图都需要拟合线性回归，而对于 Y 列，则需要线性和二次的混合。但是，在图中，所有拟合都是二次的。在某些情况下也没有拟合线。对调整代码有什么建议吗？

set.seed(14)
df <- data.frame(
  var.test = rep(c("X","Y"),each =  40),
  var.test2 = rep(c("A","B","C","D"), times = 20),
  val.test = rnorm(80,4,5),
  x = c(1:80)
)

my.formula <- c(y~x,y~x,y~x,y~x,y~x + I(x^2), y~x + I(x^2),y~x + I(x^2), y ~x)

ggplot(df, aes(x = x, y = val.test)) + 
  geom_point() +
  mapply(function(x, z) {
    data <- subset(df, var.test == z)
    list(
      geom_smooth(
        method = "glm", data = data, formula = x,
        method.args = list(family = "poisson"), color = "black"
      ),
      stat_poly_eq(formula = x, aes(label = ..eq.label..), 
                   parse = TRUE, size = 2.5, col = "black", data = data, vjust = -0.1),
      stat_poly_eq(formula = x, aes(label = ..rr.label..), 
                   parse = TRUE, size = 2.5, col = "black", data = data, vjust = 1.1)
    )
  }, my.formula, c("A", "B", "C","D","X","Y")) + 
  facet_grid(var.test2~var.test)

Answer 1

据我所知有两个问题。首先，您的回归是泊松回归，但您的数据不是计数数据 - 它包含负值，因此有时会完全失败，即使没有负值也是不合适的。我很感激这只是示例数据，因此它可能与您的实际用例无关。

其次，更重要的是对于您的实际用例，您的映射函数不太正确。您需要通过 both 分面变量进行子集化，以便您拥有 8 组独特的层 - 每个层用于 var.test 和 var.test2.[=14= 的每个交互]

ggplot(df, aes(x = x, y = val.test)) + 
  geom_point() +
  mapply(function(x, z, v) {
    data <- subset(df, var.test2 == z & var.test == v)
    list(
      geom_smooth(
        method = "glm", data = data, formula = x,
        method.args = list(family = "gaussian"), color = "black"
      ),
      stat_poly_eq(formula = x, aes(label = ..eq.label..), 
                   parse = TRUE, size = 2.5, col = "black", 
                   data = data, vjust = -0.1),
      stat_poly_eq(formula = x, aes(label = ..rr.label..), 
                   parse = TRUE, size = 2.5, col = "black",
                   data = data, vjust = 1.1)
    )
  }, 
  x = my.formula, 
  z = rep(c("A", "B", "C","D"), 2), 
  v = rep(c("X", "Y"), each = 4)) + 
  facet_grid(var.test2~var.test)

非线性拟合 ggplot2 R 中的 facet_grid 函数

Non-linear fits in facet_grid function in ggplot2 R

r

ggplot2

data-fitting