E1 244 - Detection and Estimation Theory (Jan 2022)

Assignment 2 Due Date: Feb. 20, 2022

Maximum Marks: 70

General Assignment Guidelines:

• Submission via Teams: Assignments will be allocated to you via the Teams “Assignments” feature. You
will have to upload your answer sheet via the same feature in Teams. Answer sheets sent to me or
the TAs by e-mail will not be considered. Please familiarize yourself with this feature.
• Late Submission Policy: Assignment submission beyond the deadline is allowed as per the below policy:
– Delay of 24 hours will attract 20% penalty
– Delay of 48 hours will attract 40% penalty
– Delay of 72 hours days will attract 60% penalty
– Assignments submitted beyond 72 hours will not be considered
The upload time reflected in the Teams will be considered as final. You are highly encouraged
to upload your answer sheets well before the deadline so that any potential connection/technical issues
can be resolved in time (recommended time is at least 1 hour before the deadline).
• File Logistics: You can scan your handwritten answers, or use a tablet, or typeset your answers in Latex.
Following logistics should be followed while uploading the answer sheet:
– Make sure your scans are properly visible. You can try using some scan apps (Ex. AdobeScan) to
get better results.
– The total file size should not be too large (ideally, less than 5 Mb). You can use available
apps/softwares to reduce the file size. Make sure that the answers are clearly visible while re-
ducing the size.
– Only a single pdf file is allowed to be uploaded (other formats will not be accepted).
– The order of answers in your file should be in the same as the order of questions. Check that you
have included all the pages in your file before uploading it.
– Name you file as: DET Assignmentx FirstName, where x is the assignment number (x=1,2,...)
– Mention your name, course name and submission date on the first page.
• Collaboration Policy: You are allowed to discuss the concepts/questions with your classmates. However,
the final answer should be as per your own understanding. Merely copying solutions from class-
mates/online sources will attract significant penalty and strict disciplinary action (refer to Section 13.2
and 13.3 of the IISc student information handbook - link). If you have collaborated/discussed with other
classmates or referred some online resource for a particular question, clearly mention the names of
the classmates/online resource at the beginning of your answer.

All questions carry equal marks but are not equally difficult, so you may want to strategize accordingly.

Problem 1 In this problem, we will empirically verify the ROC curve and the Bayes risk. Consider the

following hypothesis problem: H0 : Y ∼ N (0, 1) versus H1 : Y ∼ N (m, 1). Assume uniform costs and the
priors are given by P0 and P1 .
(a) Compute the Bayes test. Also specify the expressions of PD , PF and the Bayes risk J in terms of the SNR
parameter d.
(b) Consider three different values of P0 = {0.1, 0.25, 0.4} and different values of d = {0, 0.1, 0.2, · · · , 2}. For
each value of P0 and d, generate 105 random observation points Y according to the appropriate prior and
conditional distributions (you can use Matlab functions for this). Perform the Bayes test on the observations
and compute the empirical Bayes risk by averaging the risks over all the observations.
(c) Plot the theoretical Bayes risk as a continuous function of d ∈ [0, 2]. In the same figure, mark discrete
points (in different color) corresponding to the empirical Bayes risk for d = {0, 0.1, 0.2, · · · , 2}. Repeat the
same exercise in the same figure for all three values of P0 = {0.1, 0.25, 0.4}. Comment on what you observe.

1
(d) For each value of d = {1, 2, 3}, plot the theoretical ROC curve. On the same curve, mark discrete points
(in different color) corresponding to empirical values of PD and PF for each value of P0 = {0.1, 0.25, 0.4}.
Comment on what you observe.
(e) You will have to also submit the Matlab code file. The file should run without errors and generate the two
plots in parts (c) and (d). Properly label all the axes, the title of the plot and the legend of the plot. Include
these two figures in the pdf file as well.

Problem 2 Consider the following ternary (3-ary)

hypothesis testing problem:
H0 : y = w, H1 : y =
1 1 0.5
m + w, H2 : y = −m + w, where y ∈ R2 , m = , w ∼ N (0, R) where R = .
1 0.5 1
(a) Compute the minimum error probability decision rule assuming equal priors.
(b) Compute the minimum error probability.

Problem 3 Consider the following M -ary testing problem: Hk : Y ∼ N (mk , σ 2 IN ) for k = 0, 1, · · · , M − 1,

where mk ∈ RN . Assume uniform costs and equal priors.
(a) Compute the optimal Bayes decision rule for this problem.
(b) Assuming that mi and mj are orthogonal for all i, j, compute the minimum Bayes risk. You can leave the
answer in terms of an integral of Q function. Hint: Think what uncorrelatedness implies in this problem.
(c) Use the union bound to compute an upper bound to the Bayes risk.
(d) Assuming that ||mk ||2 = 2 for all k, plot the Bayes risk in part (b) and the upper bound in part (c) in
the same figure as a function of σ for σ ∈ [1, 10]. Comment on what you observe. (You may need to evaluate
some integrals numerically.)

Problem 4 Consider the following M -ary testing problem: Hk : Y ∼ Poisson((k + 1)λ), k = 0, 1, · · · , M − 1.

Assume uniform costs and equal priors.
(a) Find the optimal Bayes test.
(b) Find an expression for the boundaries of the decision region.
(c) Indicate how would you compute the probability of error.

Problem 5 Consider the following testing problem: H0 : f (y/H0 ) = 21 e−|y| versus H1 : f (y/H1 ) = e−2|y| ,
where y ∈ R. Consider the following three tests: δ1 (y) = 0 for all y, δ2 (y) = 1 for all y. The third test δ3 (y)
is a randomized test given by:

1 with probability 0.5 if y < 0,

δ3 (y) = 1 if 0 ≤ y ≤ τ,

0 if y > τ,


where τ ≥ 0 is a given constant.

(a) Compute and sketch the ROC for test δ3 .
(b) Show that a suitable randomization
√
between tests δ0 , δ1 , and δ3 improves the ROC curve for PF =
[0, α0 ] ∪ [α1 , 1], where α0 = 43 and α1 = 1 − α0 .
(c) Show that the corresponding ROC for part (b) is a straight line for PF = [0, α0 ], coincides with the ROC
of δ3 for PF = [α0 , α1 ] and is another straight line for PF = [α1 , 1].

Problem 6 Consider the following testing problem: H0 : f (y/H0 ) = (1 − β0 )β0y versus H1 : f (y/H1 ) =
(1 − β1 )β1y for y = 0, 1, 2, · · · . Let 0 < β0 < β1 < 1.
(a) Find the Bayes rule for uniform costs and equal priors.
(b) Find the minmax rule for uniform costs.
(c) Find the N-P rule with PF = α. Find PD as a function of α and sketch the ROC curve.

2
Problem 7 Consider the following testing problem: H0 : Y = W versus H0 : Y = m + W , where W is a
zero-mean multivariate Gaussian random variable with E[Wk Wl ] = σ 2 ρ|k−l| , for 1 ≤ k, l, ≤ N and |ρ| < 1.
(a) Show that the N-P test is of the form
N
X H1
bk x k ≷ η
H0
k=1
m1 y1 mk − ρmk−1 yk − ρyk−1
where b1 = , x1 = , bk = p , xk = p for k = 2, 3, · · · , N .
σ σ σ 1−ρ 2 σ 1 − ρ2
(b) Find η such that PF = α.
(c) Find PD as a function of α and sketch the ROC curve.