使用带有"gam"的geom_smooth()的等效跨度. [英] Equivalent of span using geom_smooth() with "gam"

问题描述

这可能是一个非常基本的问题，但是我还没有找到答案.当 method ="gam" 时， geom_smooth 函数中的 span 参数是否等效?我对GAM总体上并不熟悉，所以我对此表示感谢.我想为n>的数据添加更灵活的(wigglier)平滑器.1'000和 method ="loess" 需要很多时间来计算.

This is probably a very basic question but I haven't found an answer yet. Is there an equivalent to the span argument in the geom_smooth function when method = "gam"? I am not familiar with GAM's in general so I would appreciate any input on that. I want to add a more flexible (wigglier) smoother to data with n > 1'000 and method = "loess" takes a lot of time to calculate.

推荐答案

mgcv :: gam 默认情况下使用惩罚回归优化平滑度.您可以将其关闭，并使用 k 参数手动指定平滑度:

mgcv::gam by default optimizes the smoothness using penalized regression. You can switch that off and specify the smoothness manually with the k parameter:

ggplot(mpg, aes(displ, hwy)) +
  geom_point() +
  geom_smooth(method = "gam", 
              formula = y ~ s(x, bs = "cs", fx = TRUE, k = 20))

您可能应该研究软件包mgcv的文档.

You should probably study the documentation of package mgcv.

这篇关于使用带有"gam"的geom_smooth()的等效跨度.的文章就介绍到这了，希望我们推荐的答案对大家有所帮助，也希望大家多多支持IT屋！