Files
ArtStudies/M1/Monte Carlo Methods/Exercise8.rmd
2024-10-09 13:57:41 +02:00

66 lines
1.2 KiB
Plaintext

# Exercise 8 - a : Truncated Normal Distribution
```{r}
f <- function(x, b, mean, sd) {
1 / (sqrt(2 * pi * sd^2) * pnorm((mean - b) / sd)) *
exp(-(x - mean)^2 / (2 * sd^2)) *
(x >= b)
}
M <- function(b, mean, sd) {
1 / pnorm((mean - b) / sd)
}
g <- function(x, b, mean, sd) {
1 / sqrt(2 * pi * sd^2) *
exp(-(x - mean)^2 / (2 * sd^2))
}
n <- 10000
mean <- 0
sd <- 2
b <- 2
x <- numeric(0)
while (length(x) < n) {
U <- runif(1)
X <- rnorm(1, mean, sd)
x <- append(x, X[U <= (f(X, b, mean, sd) / (M(b, mean, sd) * g(X, b, mean, sd)))])
}
t <- seq(b, 7, 0.01)
hist(x, freq = FALSE, breaks = 35)
lines(t, f(t, b, mean, sd), col = "red", lwd = 2)
```
# Exercise 8 - b : Truncated Exponential Distribution
```{r}
f <- function(x, b, lambda) {
lambda * exp(-lambda * (x - b)) * (x >= b)
}
M <- function(b, lambda) {
exp(lambda * b)
}
g <- function(x, lambda) {
lambda * exp(-lambda * x)
}
n <- 10000
b <- 2
lambda <- 1
x <- numeric(0)
while (length(x) < n) {
U <- runif(1)
X <- rexp(1, lambda)
x <- append(x, X[U <= (f(X, b, lambda) / (M(b, lambda) * g(X, lambda)))])
}
t <- seq(b, 7, 0.01)
hist(x, freq = FALSE, breaks = 35)
lines(t, f(t, b, lambda), col = "red", lwd = 2)
```