如何在ggplot2中实现逻辑回归绘图？代码修改遇阻求助

Got it, let's get this sorted for you! The problem with your initial attempt is twofold: a tiny typo, and misunderstanding how ggplot handles logistic regression fits. Here's how to replicate your linear regression-style plot with a binary outcome:

Key Fixes & Explanation #

• First, Binominal is misspelled—it should be binomial (lowercase 'b'). But more importantly, logistic regression is a type of Generalized Linear Model (GLM), so you need to use method="glm" and specify the binomial family as an argument to that method, not use method="binomial" directly.

Full Code Example #

Assuming your dataset is named df, with a continuous predictor (e.g., predictor) and your binary Score1 variable (0/1), here's the code to make a plot matching your original linear regression chart's structure:

library(ggplot2)

# Recreate your plot with logistic regression fit
ggplot(df, aes(x = predictor, y = Score1)) +
  # Keep scatter points (add jitter if points overlap heavily)
  geom_point(alpha = 0.3) + 
  # Logistic regression fit with confidence intervals (matches lm's default se=TRUE)
  geom_smooth(
    method = "glm",
    method.args = list(family = binomial),
    se = TRUE,
    color = "#0072B2"  # Use your original plot's line color here
  ) +
  # Match your original plot's labels and theme
  labs(
    x = "Your Predictor Variable Name",
    y = "Probability of Score1 = 1"
  ) +
  theme_bw()  # Or use your original theme (e.g., theme_classic())

Extra Tips for Matching Your Original Chart #

• If your original linear regression plot used geom_jitter() instead of geom_point() to handle overlapping points, swap that in—binary outcomes often have overlapping 0/1 points, so jitter can help visibility.
• Copy over any other elements from your original plot: axis titles, theme settings, annotations, or color scales to keep the layout identical.
• If you need to extract the predicted probabilities manually (for labels or further analysis), you can fit the glm model separately and add predictions to your dataset, but geom_smooth() handles the plotting automatically.

内容的提问来源于stack exchange，提问作者Evan