在ggplot2中按标签拟合分组曲线

我不确定这是否是你要找的。我将数据和绘图分开并应用于stat_smooth每个组。不幸的是，平滑线不遵循配色方案。您还将看到对创建样条线的方法产生的几个警告。

plt <- ggplot(df1, aes(x=time, y=depth, colour = time)) +
    geom_label(aes(label=round(R))) + 
    scale_y_continuous(trans = "reverse") +
    geom_path(aes(group=G), size = .6, alpha = .5)

lapply(1:length(unique(df1$G)),
       function(i){
         df2 <- df1 %>% filter(G == unique(G)[i])
         plt <<- plt + 
           stat_smooth(data = df2, size = .5,
                       aes(x = time, y = depth), 
                       se = F,  method = lm, color = "darkred",
                       formula = y ~ splines::bs(x, knots = nrow(df2)))
       })

您可以使用其他参数进一步扩展它。我只是不确定你到底在期待什么。

plt <- ggplot(df1, aes(x=time, y=depth, colour = time)) +
    geom_label(aes(label=round(R))) + 
    scale_y_continuous(trans = "reverse") +
    geom_path(aes(group=G), size = .6, alpha = .5)

lapply(1:length(unique(df1$G)),
       function(i){
         df2 <- df1 %>% filter(G == unique(G)[i])
         # u <- df1 %>% {nrow(unique(.[,c(1:2)]))}
         plt <<- plt + 
           stat_smooth(
             data = df2, size = .5,
             aes(x = time, y = depth), 
             se = F,  method = lm, color = "darkred",
             formula = y ~ splines::bs(x, knots = nrow(df2),
                                       degree = ifelse(nrow(df2) <= 4,
                                                       3, nrow(df2) - 2)))
       })