Probability and Stochastic Processes (EE 591)

Assignment-1
Prof. S. Dandapat

August 3, 2022

1.
(a) Given S = {1, 2, 3}, enumerate all the fields on S.
(b) The Borel sigma field B is the minimum sigma field containing all semi-closed intervals of the form (a, b],
a, b ∈ R. Show that the open interval (a, b) ∈ B.

2. Let S = 1, 2, ... and F = Ps .

i Bool’s Inequality (Union Bound); P ( ni=1 Ai ) 6 ni=1 P (Ai )

ii Bonferroni’s inequality; P ( ni=1 Ai ) > ni=1 P (Ai ) − i,j|i<j P (Ai ∩ Aj )

3. An urn contains three balls numbered 1, 2, 3. The experiment consists of drawing a ball at random, record-
ing the number, and replacing the ball before the next ball is drawn. This is called sampling with replacement.
What is the probability of drawing the same ball twice in two tries. [Ans: 13 ]

4. An experiment consists of drawing two balls at random, with replacement from an urn containing five balls
numbered 1 to 5. The students “Dim”, “Dense”, and “Smart” were asked to compute the probability p that
2
the sum of numbers appearing on the two draws equals 5. Dim computed p = 15 , arguing that there are 15
distinguishable unordered pairs and only 2 are favorable, that is (1, 4) and (2, 3). Dense computed p = 19 ,
4
arguing that there are 9 distinguishable sums (2 to 10), of which only 1 was favorable Smart computed p = 25 ,
arguing that there were 25 distinguishable ordered outcomes of which 4 are favorable, that is, (4, 1), (3, 2),
4
(2, 3) and (1, 4). Why is p = 25 the correct answer? Explain what is wrong with the reasoning of Dense and
Dim.

5. The following problem was given to 60 students and doctors at the famous Hevardi Medical School (HMS).
Assume there exist a test to detect a disease, say D, whose prevalence is 0.001. The test has a false positive
rate of 0.005 and a correct detection rate of one. The correct detection rate is the probability that if yoy have
D, the test will say that you have D. Given that you test positive for D, what is the probability that you
actually have it? Many of the HMS experts answered 0.95 and the average answer was 0.56. Show that your
knowledge of probability is greater than that of the HMS experts by getting the right answer of 0.17. [Ans:
0.167]

1
6. First throw an unbiased die, then throw as many unbiased coins as the point shown on the die.(a) What is
the probability of obtaining 5 heads? (b) If 3 heads are obtained, what is the probability that the die showed
4?

7. A large class in probability theory is taking a multiple-choice test. For a particular question on the test,
the fraction of the examinees who know the answer is p; 1-p is the fraction that will guess. The probability
1
of answering a question correctly is unity for an examinee who knows the answer and m for a guessee; m is
the number of multiple choice alternatives. Compute the probability that an examinee knew the answer to a
mp
question given that he or she has correctly answered it. [Ans: mp+1−p ]

8. Assume there are 3 machines, A, B, and C in a semiconductor manufacturing facility that make chips. They
manufacture, respectively, 25, 35 and 40 percent of the total semiconductor chips there. Of there outputs,
respectively, 5, 4 and 2 percent of the chips are defective. A chip is drawn randomly from the combined output
of the three machines and is found defective. What is the probability that this defective chip was manufactured
by machine A? by machine B? by machine C? [Ans: (a) 0.363, (b) 0.406, (c) 0.232]

9. War-game strategists make a living by solvingproblems of the following type. There are six incoming
ballistic missiles (BM’s) against which are fired twelve antimissile missiles (AMM’s). The AMM’s are fired
so that two AMM’s are directed against each BM. The single-shot-kill probability (SSKP) of an AMM is 0.8.
The SSKP is simply the probability that an AMM can, at most, destroy only the BM against which it is fired.
Compute the probability that (a) all BM’s are destroyed, (b) at least one BM gets through to destroy the
target, and (c) exactly one BM gets through.

10. An automatic breathing apparatus (B) used in anesthesia fails with probability PB . A failure means death
to the patient unless a monitor system (M) detect the failure and alerts the physician. The monitor system
fails with probability PM . The failure of the systems are independent events. Professor X, an M.D. at Hevardi
Medical School, argues that if PM > PB installation of M is useless. Show that Prof. X needs to take on a
course on probability theory by computing the probability of a patient dying with and without the monitor
system in place. Take PM = 0.1 = 2PB .

11. Monitor three consecutive phone calls going through a telephone switching office. Classify each one as a
voice call (v) if someone is speaking, or a data call (d) if the call is carrying a modem or fax signal. Your
observation is a sequence of three letters (each letter is either v or d). For example, tow voice calls followed
by one data call corresponds to vvd. Write the elements of the following sets: (1) A1 ={first call is a voice
call}, (2) B1 ={first call is a data call}, (3) A2 ={second call is a voice call}, (4) B2 ={second call is a data
call}, (5) A3 ={all calls are the same}, (6) B3 ={voice and data alternate}, (7) A4 ={one or more voice calls},
(8) B4 ={two or more data calls}. For each pair of events A1 and B1 , A2 and B2 , and so on, identify whether
the pair of events is either mutually exclusive or collectively exhaustive or both.

12. A student’s test score T is an integer between 0 and 100 corresponding to the experimental outcomes
s0 , ..., s100 . A score of 90 to 100 is an A, 80 to 89 is a B, 70 to 79 is a C, 60 to 69 is a D, and below 60 is a

2
falling grade of F . Given that all scores between 51 and 100 are equally likely and a score of 50 or less never
occurs, find the following probabilities: (1) P [{s79 }], (2) P [{s100 }], (3) P [A], (4) P [F ], (5) P [T ≥ 80], (6)
P [T < 90], (7) P [a C grade or better], (8) P [student passes].

13. Monitor a phone call. Classify the call as voice call (V ) if someone is speaking, or a data call (D) if
the call is carrying a modem or fax signal. Classify the call as long (L) if the call lasts for more than three
minutes; otherwise classify the call as brief (B). Based on data collected by the telephone company, we use
the following probability model: P [V ] = 0.7, P [L] = 0.6, P [V L] = 0.35. Find the following probabilities:
(1)P [DL], (2)P [D ∪ L], (3)P [V B], (4)P [V ∪ L], (5)P [V ∪ D], (6)P [LB].

14. Consider a binary code with 4 bits (0 or 1) in each code word. An example of a code word is 0110.
(1) How many different code words are there?
(2) How many code words have exactly two zeros?
(3) How many code words begin with a zero?
(4) In a constant-ratio binary code, each code word has N bits. In every word, M of the N bits are 1 and the
other N − M bits are 0. How many different code words are in the code with N = 8 and M = 3?

15. Monitor three consecutive phone calls going through a telephone switching office. Classify each one as voice
call (v) if someone is speaking, or a data call (d) if the call is carrying a modem or fax signal. Your observation
is a sequence of three letters (each one is either v or d). For example, three voice calls corresponds to vvv. The
outcomes vvv and ddd have probability 0.2 whereas each of the other outcomes vvd, vdv, vdd, dvv, dvd and ddv
has probability 0.1. Count the number of voice calls NV in three calls you have observed. Consider the four
events NV = 0, NV = 1, NV = 2, NV = 3. Describe in words and also calculate the following probabilities:
(1) P [NV = 2], (2) P [NV ≥ 1], (3) P [{vvd}|NV = 2], (4) P [{ddv}|NV = 2], (5) P [NV = 2|NV ≥ 1], (6)
P [NV ≥ 1|NV = 2].

16. A department of IIT-G, provides multiple specializations in MTech, and each specialization is multidis-
ciplinary. Specialization 1 has 200 students in which 5% are from a branch (say EE). Specialization 2 has
50 students, of which 40% are from EE. Specialization 3 and 4 have 100 students with 10% EE students. A
random specialization is selected, and a student is picked.
(a) What is the probability that the selected student is from EE?
(b) A student is selected, and it is found that He/she is from EE. Based on this evidence, determine the
probability that the student is from specialization 2.
Ans: [0.1625, 0.615]

17. Data packets containing 100 bits are transmitted over a communication link. A transmitted bit is received
in error (either a 0 sent is mistaken for a 1, or a 1 sent is mistaken for a 0) with probability  = 0.01, indepen-
dent of the correctness of any other bit. The packets has been coded in such a way that if three or fewer bits
are received in error, then those bits can be corrected. If more than three bits are received received in error,
then the packet is decoded with errors.
(i) Let Sk,100−k denote the event that a received packet has k bits in error and 100 − k correctly decoded bits.
What is P [Sk,100−k ] for k = 0, 1, 2, 3?
(ii) Let C denote the event that a packet is decoded correctly. What is P [C]?

3
18. For classification tasks, machine learning offers multiple classifiers. The choice of a classifier largely de-
pends on the type of data and the underlying conditions associated with the task. Along with the classifier’s
performance, the computation cost associated with the classifier is also an important parameter for the selec-
tion of a classifier. A researcher has a task to classify ECG signals as healthy or abnormal, and he intends to
evaluate different classifiers in terms of classification performance and computation cost. For the given naive
Bayes classifier, the computation cost can be High (H) or Low (L), and the performance is either Good (G)
or Poor (P ). The experiment has following informations: P [G] = 0.5, P [HG] = 0.2, P [LP ] = 0.1. What is
the sample space of the experiment? Calculate the following probabilities: (a) P [P ], (b) P [LG], (c) P [H]
Ans: [(a) 0.5, (b) 0.3, (c) 0.6]

19. A researcher is working on building an algorithm to extract clinically relevant information from radiology
reports. The system used to build the algorithm has 16 GB of RAM and two GPUs of 6GB and 4GB. Based
on the computational complexity, the program can use either only the RAM (R), or RAM and one GPU (G1),
or RAM and both the GPUs (G2). The program may run for a couple of hours. Consider long (L) if it
runs for more than 2 hours else brief (B). Table 1 describes the probabilities of possible cases. What is the
probability P [R] that the program runs only on RAM? What is the probability a program is brief? What is
the probability that a program is long or two GPUs have been used?
Ans: [0.5, 0.6, 0.5]

Table 1: Probabilities.

H0 H1 H2
L 0.1 0.1 0.2
B 0.4 0.1 0.1

20. Coronavirus disease (COVID-19) causes severe acute respiratory syndrome in affected people. People who
have underlying medical conditions, such as hypertension and diabetes, have shown worse prognosis. In a
study of critical patients of COVID-19, it was found that 16% had hypertension, 10% had diabetes, and that
10% of critical patients that had either hypertension or diabetes had both the conditions. (a) What is the
probability of P [HD] that a critical patient has both hypertension (H) and diabetes (D)? (b) What is the
conditional probability that a patient has hypertension, given that it has diabetes?
Ans: [(a) 0 .0236, (b) 0.1475]