在R中生成iid二元变量X=(x1,x2)的技术求助

Got it, let's break down how to create this bivariate variable and plot it in R step by step. First, let's align on the requirements: we need 1000 independent, identically distributed (iid) observations where x1 follows a standard normal distribution, and x2 is approximately equal to x1 when x1 is between -1 and 1, exactly equal to x1 otherwise.

Step 1: Generate the Standard Normal x1 #

First, we'll create x1 using R's built-in normal distribution function. Setting a random seed ensures your results are reproducible (you can change the seed value or remove this line if you don't need consistency).

# Set seed for reproducible results
set.seed(123)

# Generate 1000 standard normal observations for x1
sample_size <- 1000
x1 <- rnorm(sample_size, mean = 0, sd = 1)

Step 2: Construct x2 Based on x1 #

You mentioned x2 is "approximately equal to x1" when -1 ≤ x1 ≤ 1. A common way to implement this is adding a small amount of random noise to x1 in that interval (you can adjust the noise level to match your definition of "approximate"). For values outside [-1, 1], we just use x1 directly.

# Create x2: add small normal noise to x1 in [-1,1], use x1 otherwise
x2 <- ifelse(abs(x1) <= 1,
             x1 + rnorm(sample_size, mean = 0, sd = 0.1),  # Small noise for "approximate"
             x1)

Customization Tip: #

If your "approximate" means something else (e.g., uniform noise instead of normal, or fixed rounding), tweak the noise line:

• For uniform noise between -0.1 and 0.1: x1 + runif(sample_size, -0.1, 0.1)
• For exact equality (if "approximate" was a typo): just use x1 instead of the noise line

Step 3: Visualize the Bivariate Data #

We'll cover two common plotting approaches: base R (quick and simple) and ggplot2 (customizable and polished).

Option 1: Base R Scatter Plot #

Great for fast exploration:

# Scatter plot with transparency to avoid overlapping points
plot(x1, x2,
     pch = 16,  # Solid circle points
     col = adjustcolor("darkblue", alpha.f = 0.5),  # Semi-transparent blue
     main = "Scatter Plot of X=(x1, x2)",
     xlab = "x1 ~ N(0,1)",
     ylab = "x2")

# Add dashed red lines to mark the [-1, 1] interval for x1
abline(v = c(-1, 1), lty = 2, col = "red")

Option 2: ggplot2 (Tidyverse Style) #

If you prefer more control over aesthetics, use ggplot2 (install it first with install.packages("ggplot2") if you haven't):

library(ggplot2)

# Convert data to a data frame (required for ggplot)
bivariate_data <- data.frame(x1 = x1, x2 = x2)

ggplot(bivariate_data, aes(x = x1, y = x2)) +
  geom_point(alpha = 0.5, color = "darkblue") +  # Semi-transparent points
  geom_vline(xintercept = c(-1, 1), linetype = "dashed", color = "red") +  # Interval markers
  labs(title = "Scatter Plot of Bivariate Variable X=(x1, x2)",
       x = "x1 ~ N(0,1)",
       y = "x2") +
  theme_minimal()

Quick Check #

To verify your data matches expectations, run these lines to summarize key stats:

# Summary stats for x1 and x2
summary(x1)
summary(x2)

# Check how many x1 values fall in [-1,1]
sum(abs(x1) <= 1) / sample_size  # Should be ~68% (empirical rule for normal distribution)

内容的提问来源于stack exchange，提问作者Derek