# Metropolis-Hastings Algorithm

- Suppose that the number of children per household in the U.S. has the following distribution

\[\begin{align*} x & & 0 & & 1 & & 2 & & 3 & & 4 & & 5\\ \text{P}(X=x) & & 0.15 & & 0.18 & & 0.35& & 0.20 & & 0.08 & &0.04 \end{align*}\]

- Describe in detail how you could implement by hand a Metropolis-Hastings algorithm to simulate from this distribution, using a random walk to neighboring states as the proposal chain.
- Write a few lines of code to implement and run your algorithm, and summarize the approximate distribution. Does your algorithm seem to work?
- Specify the transition matrix for the M-H algorithm in the previous part.
- Find the stationary distribution of the M-H chain. Is it the target distribution?

- The standard double exponential distribution has a pdf which satisfies \[ f(x) \propto e^{-|x|}, \qquad -\infty<x<\infty \]

- Describe in detail a Metropolis-Hastings algorithm for simulating from this distribution.
- Write a few lines of code to implement the algorithm and run it. Does the algorithm seem to produce values from the target distribution?