


Study with the several resources on Docsity
Earn points by helping other students or get them with a premium plan
Prepare for your exams
Study with the several resources on Docsity
Earn points to download
Earn points by helping other students or get them with a premium plan
The question booklet for the midterm 2 exam of ee 126, a university course taught by professor martin wainwright in spring 2006. The exam consists of three problems related to probability theory and statistics, including calculating probabilities, expected values, and moment generating functions. Students are required to solve the problems using the provided question booklet and a calculator, and are allowed to bring handwritten notes. The document also includes instructions for the exam, such as the duration, format, and rules.
Typology: Exams
1 / 4
This page cannot be seen from the preview
Don't miss anything!



3
or ∑ 5 k=0(1/2) k (^) are also fine.
Problem 1: (12 points)
Diana the Daredevil is trying to break the world land-speed record using her rocket- powered motorcycle. In order to do so, she needs to cover more than 500 yards in 10 seconds. If her motorcycle has Z pounds of rocket fuel to start, then it travels a distance of X = 50
Z yards in 10 seconds.
(a)(5 points) Suppose that the amount of rocket fuel Z is a random variable uniformly distributed over [50, 150]. Compute the PDF and CDF of the distance X.
(b)(2 points) Compute the probability that Diana breaks the world record on any given trial.
(c)(5 points) Now suppose that Diana is allowed to take a total of n trials in order to try and break the world record. (Here n is some fixed positive integer: i.e., n ∈ { 1 , 2 , 3 ,.. .}). The amount of rocket fuel at the start is independent from trial to trial. What is the smallest possible integer n such that Diana has a better than 90% chance of breaking the world record over the n trials?
Problem 3: (14 points) George the Gambler is a good poker player: every round that he plays, he wins a random amount Z of money, distributed as Z ∼ N (μ, σ^2 ) with μ > 0. Every night, George goes to the local poker club, and plays T = 1 + V rounds of poker, where V is a Poisson random variable with parameter λ = 5. Let X be the total amount of money that George wins in a given night.
(a)(2 points) In the absence of any further information, what is the best estimate of X? (Here “best” is measured by the minimum mean-squared error.)
(b)(3 points) Suppose that you observe that George plays T = t rounds of poker (where t is some positive integer). Now what is the best estimate of X (again measured in terms of minimim mean-squared error)?
(c)(5 points) Now suppose that by peeking into George’s wallet at the end of the night, you make a noisy observation Y of the amount of money X that he won—say of the form
Y = X + W
where W ∼ N (0, 1) is Gaussian noise independent of X. Compute the Bayes’ least squares estimate of X based on observing T = t and Y = y. Also compute the linear least squares estimate (LLSE) of X based on observing T = t and Y = y, and the error variance of the LLSE.
(d)(4 points) Now suppose that you observe that {T ≤ 2 } and Y = y. Compute the Bayes’ least squares estimate of X based on this information.