CODEWITHSUNDEEP
rggplot2beta
David
David

Reputation: 11

Plot for beta(1/2, 1/2) distribution

I'm trying to adjust a plot of the beta(1/2, 1/2) distribution to fit into the vertical bars between x=.05 and 1.0. The code below gets very close.

library(ggplot2)

# Set the shape parameters
shape1 <- .5
shape2 <- .5

# Set the lower and upper limits
lower <- 0.05
upper <- 0.99888

# Calculate the probability density function (PDF) at every point between 0.10 and 1
df <- data.frame(x = seq(lower, upper, length.out = 1000), y = dbeta(seq(lower, upper, length.out = 1000), shape1, shape2))

# Plot the PDF with custom tick marks and labels on the x-axis, and a tick mark at (0,0)
ggplot(df, aes(x = x, y = y)) + 
  geom_line() +
  xlab(expression(kappa[j])) + ylab("") +
  scale_x_continuous(limits = c(0, 1), expand = c(0, 0), 
        breaks = seq(0, 1, 0.5), labels = c("0.0", "0.5", "1.0")) +
  theme_classic() +
  theme(axis.title.y = element_blank(), axis.text.y = element_blank(), axis.ticks.y = element_blank()) +
  #geom_vline(xintercept = 0, color = "black") +
  geom_vline(xintercept = 0.05, color = "black") +
  geom_vline(xintercept = 1, color = "black") +
  labs(title = "")

Here is the plot that is generated by this code.

enter image description here

I need the plot to fit between x=.05 and x=1 and maintain the U shape of the distribution. I realize that this will make it a bit off center from 0.5, but that's the point.

Thanks in advance,

David

Upvotes: 0

Views: 260

Answers (1)

Harrison Jones
Harrison Jones

Reputation: 2506

As mentioned in my comment, this is artificially removing instances of the beta distribution that are below 0.05.

df <- data.frame(x = rbeta(n = 100000, 0.5, 0.7))

df %>%
  filter(x > 0.05) %>%
  ggplot(aes(x = x)) +
  geom_histogram(bins = 100) +
  geom_vline(xintercept = 0.05, color = "black", linetype = "longdash") +
  geom_vline(xintercept = 1, color = "black", linetype = "longdash")

Could be converted into a line graph if needed, but this was easier to generate quickly.

Upvotes: 0

Related Questions