在ggplot中围绕候选点绘制半径0.1的圆形是否可行

当然可以实现，你可以借助ggforce包中的geom_circle()函数来绘制指定半径的圆周，步骤如下：

1. 首先安装并加载ggforce包（如果尚未安装）：

install.packages("ggforce")
library(ggforce)

2. 在你的现有绘图代码中添加geom_circle()图层，针对candidates数据框中的两个候选点绘制圆周：

修改后的完整绘图代码如下：

ggplot(df, aes(lambda, log_likelihood)) +
    geom_line() +
    geom_point(aes(x = lambda_mle, y = log_likelihood[which.min(abs(lambda_values - lambda_mle))], color = "Estimated Value"), size = 3) +
    geom_point(aes(x = 5, y = log_likelihood[which.min(abs(lambda_values - 5))], color = "True Value"), size = 3) +
    geom_point(data = candidates, aes(x = lambda, y = log_likelihood, color = candidate), size = 3) +
    # 添加候选点的圆周
    geom_circle(data = candidates, 
                aes(x0 = lambda, y0 = log_likelihood, r = 0.1, color = candidate),
                linetype = "solid", # 可改为"dashed"使用虚线
                fill = NA) + # 设为NA表示只画圆周，不填充内部
    labs(x = "Lambda", y = "Log-Likelihood", 
         title = "Log-Likelihood Function for Poisson Distribution") +
    theme_minimal() +
    scale_color_manual(values = c("True Value" = "green", "Estimated Value" = "red", "Candidate Solution #1" = "purple", "Candidate Solution #2" = "orange")) +
    geom_text(aes(lambda_mle, log_likelihood[which.min(abs(lambda_values - lambda_mle))], 
                  label = round(lambda_mle, 2), color = "Estimated Value"), hjust = -0.2, vjust = -1.5) +
    geom_text(aes(5, log_likelihood[which.min(abs(lambda_values - 5))], 
                  label = 5, color = "True Value"), hjust = -0.2, vjust = 1.5) +
    geom_text(data = candidates, aes(x = lambda, y = log_likelihood, label = lambda, color = candidate), hjust = 0.5, vjust = -1) +
    xlim(5, 5.5) +
    labs(color = "Legend")

关键说明： #

• geom_circle()的x0和y0参数对应圆心坐标（即候选点的lambda和对数似然值），r参数设置半径为0.1；
• fill = NA确保只绘制圆周轮廓，不会填充内部区域；
• linetype参数可根据需求设置为solid（实线）或dashed（虚线）；
• 颜色映射沿用了原有的候选点颜色，保证圆周和对应点颜色一致。

如果不想额外安装包，也可以用annotate()手动添加两个圆周，但这种方式需要逐个指定坐标，不如用geom_circle()结合数据框灵活：

# 手动添加圆周的示例（无需ggforce）
annotate("circle", xmin = 5.1 - 0.1, xmax = 5.1 + 0.1, ymin = candidates$log_likelihood[1] - 0.1, ymax = candidates$log_likelihood[1] + 0.1, color = "purple", fill = NA) +
annotate("circle", xmin = 5.2 - 0.1, xmax = 5.2 + 0.1, ymin = candidates$log_likelihood[2] - 0.1, ymax = candidates$log_likelihood[2] + 0.1, color = "orange", fill = NA)

内容的提问来源于stack exchange，提问作者stats_noob