如何在R语言norm_f()函数中集成log2转换可视化并生成2×2正态分布评估图

Got it! Here's a refined version of your norm_f() function that generates a 2×2 plot grid—showing normality assessments for both the original variable and its log2-transformed counterpart. I've also added a safety check to avoid errors with non-positive values (since log2 can't handle those):

norm_f <- function(z, hist. = list(col = "lightgray"), lines. = list(lwd = 2), curve. = list(col = "red", lwd = 2), qqnorm. = list(col = "black", pch = 1, cex = 1.2), qqline. = list(lwd = 2)) {
  zname <- deparse(substitute(z))
  
  # Prevent log2 transformation errors for non-positive values
  if (any(z <= 0)) {
    stop(paste("Can't log2-transform", zname, "- contains values ≤ 0"))
  }
  z_log2 <- log2(z)
  z_log2_label <- paste("log2(", zname, ")", sep = "")
  
  # Set up 2×2 plot layout
  par(mfrow = c(2, 2), mar = c(5, 5, 2, 2))
  
  # Reusable helper to plot normality for a single variable
  plot_single_norm <- function(var, var_label) {
    var_density <- density(var)
    hist_params <- hist.
    
    # Auto-set ylim if not provided
    if (!"ylim" %in% names(hist_params)) {
      y_max <- max(var_density$y, dnorm(seq(min(var), max(var), len = 21), mean = mean(var), sd = sd(var)))
      hist_params$ylim <- c(0, y_max)
    }
    
    # Auto-set plot labels if not provided
    if (!"main" %in% names(hist_params)) hist_params$main <- paste("Histogram of", var_label)
    if (!"xlab" %in% names(hist_params)) hist_params$xlab <- var_label
    
    # Generate histogram + density + normal curve
    do.call(hist, c(list(var, prob = TRUE), hist_params))
    do.call(lines, c(list(var_density), lines.))
    do.call(curve, c(list(substitute(dnorm(x, mean=mean(var), sd=sd(var))), add = TRUE), curve.))
    
    # Generate QQ plot + QQ line
    do.call(qqnorm, c(list(var), qqnorm.))
    do.call(qqline, c(list(var), qqline.))
  }
  
  # Plot original variable
  plot_single_norm(z, zname)
  # Plot log2-transformed variable
  plot_single_norm(z_log2, z_log2_label)
}

Key Improvements: #

• Safety Check: Blocks log2 transformation for variables with non-positive values to avoid runtime errors.
• Code Reusability: Added a plot_single_norm() helper function to eliminate duplicate code for original and transformed variables.
• Auto-Labeling: Automatically creates descriptive titles/labels for the log2-transformed variable (e.g., log2(Sepal.Length) for your iris example).
• 2×2 Layout: Adjusted par(mfrow=c(2,2)) to fit all four plots neatly.

Example Usage: #

# Test with iris dataset's Sepal.Length (all values are positive, so log2 works)
norm_f(iris$Sepal.Length)

This will produce:

• Top-left: Histogram + density curve + normal curve for original Sepal.Length
• Top-right: QQ plot for original Sepal.Length
• Bottom-left: Histogram + density curve + normal curve for log2(Sepal.Length)
• Bottom-right: QQ plot for log2(Sepal.Length)

内容的提问来源于stack exchange，提问作者cmirian