Scribd
Upload
13 views

Problem Set 1

This document is a problem set for a detection and estimation theory course. It contains 8 problems related to concave functions, continuous and differentiable functions, minimum Bayes risk, and generating Gaussian random variables from uniform random variables. Students are encouraged to collaborate but must write solutions individually without assistance. The problem set is due on January 27th.

Uploaded by

MelbinGeorge
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
13 views

Problem Set 1

This document is a problem set for a detection and estimation theory course. It contains 8 problems related to concave functions, continuous and differentiable functions, minimum Bayes risk, and generating Gaussian random variables from uniform random variables. Students are encouraged to collaborate but must write solutions individually without assistance. The problem set is due on January 27th.

Uploaded by

MelbinGeorge
Copyright
© © All Rights Reserved
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 1

E1–244 Detection and Estimation Theory

Discussions: 20 January 2015


Due: Thursday 27 January 2015
Problem Set 1
Instructor: Rajesh Sundaresan TA: None

Remarks:

• Collaboration, discussion, and working in teams to solve problems is strongly encouraged.

• To test your understanding, write the solution to each problem in your own words without
referring to a friend, text, or class notes.

• You can solve the programming problems in teams of three. One submission per team
suffices. Indicate team members.

Problems:

1. Show that for any index set ∆, the function V : [0, 1] → R given by

x 7→ V (x) = inf [aδ x + bδ ]


δ∈∆

is a concave function.

2. Let V : [0, 1] → R be a concave function. Show the following.

• f is continuous in the open set (0, 1).


• Left and right derivatives exist at every point in (0, 1).

3. Show that the minimum Bayes risk V (π0 ) is continuous at the end-points 0 and 1, that
it has a right derivative at 0 and a left derivative at 1. Find the slopes. (Hint: Write out
V (π0 ). For the right derivative at 0, use p1 (y) ≤ τ (π0 )p0 (y) for the decision region for
H0 , where τ (π0 ) is O(π0 ) as π0 approaches 0).

4. Problem 2 in Section II.F.

5. Problem 11 in Section II.F.

6. Problem 16 in Section II.F.

7. Write a program in matlab to generate a Gaussian random variable with mean 0 and
variance 1 using only the function rand that generates a uniform random variable between
[0, 1]. (Hints: It is easier to generate two such Gaussians. What is the distribution of
R2 = X12 + X22 ? Of the angle θ = arctan(X2 /X1 )? How will you generate samples for
R and Θ using rand? Transform the generated samples into samples for X1 and X2 .
Vectorise your code.)

8. Let d = (µ1 −µ0 )/σ in the location testing problem with Gaussian errors. Plot the receiver
operating characteristics for d = 0, d = ∞ and three other intermediate values of d. (You
can use erf or erfc or build your Q function.)

Problem Set 1-1

You might also like