R 绘制具有平滑效果的图形

Question

我有一个示例 data.table 如下

dt <- structure(list(x = c(-5.08454829941579, -5.03954101414227, -4.96110744424707, 
-3.91467506637747, -3.08288789758443, -2.10944056281588, -1.35385839597117, 
-0.991302108693737, -0.747580793596581, 0.021222777307317, 1.04703618291297, 
1.71724303836327, 1, 0), y = c(38.7610606189486, 39.1292360562013, 
25.6794493857896, 29.2344899886708, 30.0047839165718, 34.1031239345073, 
40.5743610428116, 43.4840735782065, 34.863487579448, 33.6026425302808, 
41.2960346180424, 42.0262260478778, 47.1194103133183, 59.6173834356779
), seq = 1:14), row.names = c(NA, -14L), class = c("data.table", 
"data.frame"))
> dt
              x        y seq
 1: -5.08454830 38.76106   1
 2: -5.03954101 39.12924   2
 3: -4.96110744 25.67945   3
 4: -3.91467507 29.23449   4
 5: -3.08288790 30.00478   5
 6: -2.10944056 34.10312   6
 7: -1.35385840 40.57436   7
 8: -0.99130211 43.48407   8
 9: -0.74758079 34.86349   9
10:  0.02122278 33.60264  10
11:  1.04703618 41.29603  11
12:  1.71724304 42.02623  12
13:  1.00000000 47.11941  13
14:  0.00000000 59.61738  14

我使用以下代码创建了一个图表 -

library(ggplot2)
library(data.table)
dt %>% ggplot(aes(x, y)) + 
  geom_point() + geom_path(alpha = 0.25) +
  stat_smooth(method = lm, formula = y ~ poly(x, 5), se = FALSE) 

有没有办法确保平滑的线按照seq列中的顺序？我有一些点应该构成一个完整的圆圈的情况。

另外，如何增加平滑线的平滑度，使其更接近图上的点？

谢谢！

Answer 1

看起来您不想要stat/geom_smooth()，因为我认为您不想回归y ~ x（这是统计数据所做的），您只想要一条平滑的线。我想建议改用 ggforce::geom_bspline()。下面的例子：

library(ggplot2)
library(ggforce)

dt <- structure(list(
  x = c(-5.08454829941579, -5.03954101414227, -4.96110744424707, 
        -3.91467506637747, -3.08288789758443, -2.10944056281588, -1.35385839597117, 
        -0.991302108693737, -0.747580793596581, 0.021222777307317, 1.04703618291297, 
        1.71724303836327, 1, 0), 
  y = c(38.7610606189486, 39.1292360562013, 
        25.6794493857896, 29.2344899886708, 30.0047839165718, 34.1031239345073, 
        40.5743610428116, 43.4840735782065, 34.863487579448, 33.6026425302808, 
        41.2960346180424, 42.0262260478778, 47.1194103133183, 59.6173834356779
  ), seq = 1:14), 
  row.names = c(NA, -14L), class = c("data.table", "data.frame"))

ggplot(dt, aes(x, y)) +
  geom_point() +
  geom_bspline()

^{由 reprex package (v0.3.0) 于 2021 年 3 月 12 日创建}