Coin flipping: MLE and MAP. (10 points) Suppose you observed coin flips consisting of NH heads and NT tails. You model the coin flips as independent Bernoulli random variables with a common parameter ̘ that represents the probability of heads. In this problem, you'll estimate ̘ in two different ways.
Maximum Likelihood
1. Write the likelihood,
ℒ(̘) = p(X|̘) = ̣ (NH + NT, i=1) p(xi|̘).
The result should be a formula involving ̘, NH, and NT.
2. Find the parameter ̘MLE that maximizes the likelihood,
̘MLE = arg max ℒ(̘).
(Hint: take the derivative and set it to zero).
3. What is ̘MLE when NH = 3 and NT = 2? When NH = 30 and NT = 20? (Hint: use the expression you found in (2).)