University of Toronto Scarborough Department of Computer & Mathematical Sciences STAB52H3 Introduction to Probability Term Test 2 November 16, 2020 Duration: 60 minutes Examination aids allowed: Open notes-books Instructions: • Read the questions carefully and answer only what is being asked. • Answer all questions directly on the examination paper; use the last pages if you need more space, and provide clear pointers to your work. • Show your intermediate work, and write clearly and legibly. Question: 1 2 3 4 5 6 7 8 9 10 Total Points: 20 20 20 20 20 20 20 20 20 20 200 Score: 1. (20 points) Consider two random variables X and Y that are binary valued, i.e., the set of possible values is {0, 1}. Let the marginal distribution of X be P(X = 0) = p, P(X = 1) = 1− p. It is given that the conditional distribution of Y given X is P(Y = 0|X = 0) = 3 4 , P(Y = 1|X = 0) = 1 4 , P(Y = 1|X = 1) = 1 2 , P(Y = 0|X = 1) = 1 2 . (a) (10 points) Find the value of p ∈ [0, 1] such that the marginal distributions of X and Y are the same. (b) (10 points) For the value of p given in (a), find the conditional distribution of X given Y . Solution: (a) By the law of total probability, we have P(Y = 0) = P(Y = 0|X = 0)P(X = 0) + P(Y = 0|X = 0)P(X = 0) = 3 4 p+ 1 2 (1− p). We require that X and Y have the same marginal distribution. So P(Y = 0) = 3 4 p+ 1 2 (1− p) = p = P(X = 0). Solving gives p = 2 3 . Since X and Y are binary valued, if p = 2 3 then they have the same marginal distributions. (b) We have P(X = 0|Y = 0) = P(Y = 0|X = 0)P(X = 0) P(Y = 0) = P(Y = 0|X = 0) = 3 4 , P(X = 1|Y = 0) = 1− P(X = 0|Y = 0) = 1 4 , P(X = 1|Y = 1) = P(Y = 1|X = 1)P(X = 1) P(Y = 1) = P(Y = 1|X = 1) = 1 2 , P(X = 0|Y = 1) = 1− P(X = 1|Y = 1) = 1 2 . (Note that the transition probabilities from X to Y are the same as those from Y to X.) Page 2 of 10 (contd...) 2. (20 points) Consider two random variables X and Y that are binary valued, i.e., the set of possible values is {0, 1}. Let the marginal distribution of X be P(X = 0) = p, P(X = 1) = 1− p. It is given that the conditional distribution of Y given X is P(Y = 0|X = 0) = 2 3 , P(Y = 1|X = 0) = 1 3 , P(Y = 1|X = 1) = 1 2 , P(Y = 0|X = 1) = 1 2 . (a) (10 points) Find the value of p ∈ [0, 1] such that the marginal distributions of X and Y are the same. (b) (10 points) For the value of p given in (a), find the conditional distribution of X given Y . Solution: (a) By the law of total probability, we have P(Y = 0) = P(Y = 0|X = 0)P(X = 0) + P(Y = 0|X = 0)P(X = 0) = 2 3 p+ 1 2 (1− p). We require that X and Y have the same marginal distribution. So P(Y = 0) = 2 3 p+ 1 2 (1− p) = p = P(X = 0). Solving gives p = 3 5 . Since X and Y are binary valued, if p = 3 5 then they have the same marginal distributions. (b) We have P(X = 0|Y = 0) = P(Y = 0|X = 0)P(X = 0) P(X = 0) = P(Y = 0|X = 0) = 2 3 , P(X = 1|Y = 0) = 1− P(X = 0|Y = 0) = 1 3 , P(X = 1|Y = 1) = P(Y = 1|X = 1)P(X = 1) P(Y = 1) = P(Y = 1|X = 1) = 1 2 , P(X = 0|Y = 1) = 1− P(X = 1|Y = 1) = 1 2 . (Note that the transition probabilities from X to Y are the same as those from Y to X.) Page 3 of 10 (contd...) 3. (20 points) Let (X, Y ) be a random vector whose distribution is given by P((X, Y ) = (0, 0)) = P((X, Y ) = (0, 1)) = P((X, Y ) = (1, 1)) = 1 3 . (a) (4 points) Find the marginal distributions of X and Y , and give their name and parameters. (b) (16 points) Find the correlation coefficient Cor(X, Y ) = ρXY between X and Y . Solution: (a) We have X ∼ Bernoulli(1 3 ) and Y ∼ Bernoulli(2 3 ). (b) The covariance is given by Cov(X, Y ) = E[XY ]− E[X]E[Y ]. We have E[XY ] = 1 3 0 · 0 + 1 3 0 · 1 + 1 3 1 · 1 = 1 3 , Recall that if Z ∼ Bernoulli(p) then E[Z] = p and Var(Z) = p(1− p). So E[X] = 1 3 , E[Y ] = 2 3 , and we have Cov(X, Y ) = 1 3 − 1 3 2 3 = 1 9 . Also Var(X) = 1 2 2 3 = 2 9 , Var(Y ) = 2 3 1 3 = 2 9 . Thus Cor(X, Y ) = Cov(X, Y )√ Var(X) √ Var(Y ) = 1 2 . 4. (20 points) Let (X, Y ) be a random vector whose distribution is given by P((X, Y ) = (0, 1)) = P((X, Y ) = (1, 0)) = P((X, Y ) = (1, 1)) = 1 3 . (a) (4 points) Find the marginal distributions of X and Y , and give their name and parameters. (b) (16 points) Find Var(X + Y ). Page 4 of 10 (contd...) Solution: (a) We have X ∼ Bernoulli(2 3 ) and Y ∼ Bernoulli(2 3 ). (b) We have Var(X + Y ) = Var(X) + 2Cov(X, Y ) + Var(Y ). The covariance is given by Cov(X, Y ) = E[XY ]− E[X]E[Y ]. We have E[XY ] = 1 3 0 · 1 + 1 3 1 · 0 + 1 3 1 · 1 = 1 3 , Recall that if Z ∼ Bernoulli(p) then E[Z] = p and Var(Z) = p(1− p). So E[X] = 2 3 , E[Y ] = 2 3 , and we have Cov(X, Y ) = 1 3 − 2 3 2 3 = −1 9 . Also Var(X) = Var(Y ) = 2 3 1 3 = 2 9 . Thus Var(X + Y ) = 2 9 − 21 9 + 2 9 = 2 9 . 5. (20 points) Consider two independent Exponential RVs: X ∼ Exponential(1) and Y ∼ Exponential(2). Find the probability P(X > Y ). Solution: The joint PDF is fX,Y (x, y) = fX(x)fY (y) = e −x2e−2y, ∀x, y > 0. The probability is given by the integral of the joint PDF over the region R = {(x, y) ∈ R : x > y} = Page 5 of 10 (contd...) {(x, y) ∈ R : 0 < x, 0 < y < x} P(X > Y ) = ∫∫ R fX,Y (x, y)dxdy = ∫ ∞ x=0 [∫ x y=0 2e−xe−2ydy ] dx = ∫ ∞ x=0 e−x [∫ x y=0 2e−2ydy ] dx = ∫ ∞ x=0 e−x [−e−2y]x 0 dx = ∫ ∞ x=0 e−x [ 1− e−2x] dx = ∫ ∞ x=0 e−xdx− ∫ ∞ x=0 e−3xdx = 1− [−1 3 e−3x ]∞ x=0 = 1− [0 + 1 3 ] = 2 3 . 6. (20 points) Consider two independent and identically distributed Exponential RVsX, Y ∼ Exponential(1). Find the probability P(X > 2Y ). Solution: The joint PDF is fX,Y (x, y) = fX(x)fY (y) = e −xe−y, ∀x, y > 0. The probability is given by the integral of the joint PDF over the region R = {(x, y) ∈ R : x > 2y} = Page 6 of 10 (contd...) {(x, y) ∈ R : 0 < x, 0 < y < x/2} P(X > 2Y ) = ∫∫ R fX,Y (x, y)dxdy = ∫ ∞ x=0 [∫ x y=0 e−xe−ydy ] dx = ∫ ∞ x=0 e−x [∫ x/2 y=0 e−ydy ] dx = ∫ ∞ x=0 e−x [−e−y]x/2 0 dx = ∫ ∞ x=0 e−x [ 1− e−x/2] dx = ∫ ∞ x=0 e−xdx− ∫ ∞ x=0 e−3x/2dx = 1− [−2 3 e−3x/2 ]∞ x=0 = 1− [0 + 2 3 ] = 1 3 . 7. Let the RV X follow Uniform(0, 1) distribution, and define the new RV Y = 1 1+X . (a) (3 points) What is the range of possible values of Y ? (b) (10 points) Find the CDF of Y . (c) (7 points) Find the PDF of Y . Solution: (a) Since X ranges in [0, 1], then Y = 1 1+X will take values between 1 1+0 = 1 and 1 1+1 = 1 2 , i.e. Y ∈ [1 2 , 1] (b) The CDF of the Uniform(0,1) is the identity function, FX(x) = x, ∀x ∈ (0, 1). We have FY (y) = P(Y ≤ y) = P ( 1 1 +X ≤ y ) = P ( 1 +X ≥ 1 y ) = P ( X ≥ 1 y − 1 ) = 1− P ( 1 +X < 1 y ) = 1− FX ( 1 y − 1 ) = 1− ( 1 y − 1 ) = 2− 1 y , ∀y ∈ [1 2 , 1] Page 7 of 10 (contd...) (c) There are two ways to find the PDF of Y : a) Differentiate the CDF of Y form the previous part: fY (y) = d dy FY (y) = d dy (2− y−1) = y−2 = 1 y2 , ∀y ∈ [1 2 , 1] b) Use the PDF method: h(x) = (1 + x)−1 ⇒ { h−1(y) = y−1 − 1 h′(x) = −(1 + x)−2 fY (y) = fX(h −1(y)) |h′(h−1(y))| = 1 | − (1 + h−1(y))−2| = 1 (1 + y−1 − 1)−2 = 1 y2 , ∀y ∈ [1 2 , 1] 8. Let the RV X follow Uniform(0, 1) distribution, and define the new RV Y = 1−X2. (a) (3 points) What is the range of possible values of Y ? (b) (10 points) Find the CDF of Y . (c) (7 points) Find the PDF of Y . Solution: (a) Since X ranges in [0, 1], then Y = 1 −X2 will take values between 1 − 02 = 1 and 1− 12 = 0, i.e. Y ∈ [0, 1] (b) The CDF of the Uniform(0,1) is the identity function, FX(x) = x, ∀x ∈ (0, 1). We have FY (y) = P(Y ≤ y) = P ( 1−X2 ≤ y) = P ( X2 ≥ 1− y) = P(X ≤√1− y) = 1− P ( X < √ 1− y ) = 1− FX (√ 1− y ) = 1− √ 1− y, ∀y ∈ [0, 1] (c) There are two ways to find the PDF of Y : a) Differentiate the CDF of Y form the previous part: fY (y) = d dy FY (y) = d dy ( 1− √ 1− y ) = d dy ( 1− (1− y)12 ) = 1 2 (1− y)−12 = 1 2 √ 1− y , ∀y ∈ [0, 1] Page 8 of 10 (contd...) b) Use the PDF method: h(x) = 1− x2 ⇒ { h−1(y) = √ 1− y h′(x) = −2x fY (y) = fX(h −1(y)) |h′(h−1(y))| = 1 | − 2(√1− y)| = 1 2( √ 1− y) , ∀y ∈ [0, 1] 9. Consider two RVs X, Y where X has marginal PDF fX(x) = { x/2, x ∈ (0, 2) 0, otherwise and Y has conditional PDF fY |X(y|x) = { 1/x, y ∈ (0, x) 0, otherwise (i.e. Y given X is uniformly distributed in (0, X).) (a) (10 points) Find the marginal PDF of Y ; are X and Y independent? (justify your answer). (b) (10 points) Find the conditional PDF of X given Y , and identify its distribution (i.e. give its name and parameters). Solution: (a) The joint PDF of X, Y is given by fX,Y (x, y) = fX(x)fY |X(y|x) = { x 2x = 1 2 , 0 < y < x < 2 0, otherwise The marginal PDF of Y is found by integrating out X from the joint PDF fY (y) = ∫ ∞ x=−∞ fX,Y (x, y)dx = {∫ 2 x=y 1 2 dx = 2−y 2 , y ∈ (0, 2) 0, otherwise The RVs are NOT independent, since the conditional and marginal PDFs of Y are different (moreover, the conditional range of Y depends on X). (b) The conditional PDF of X given Y is fX|Y (x|y) = fX,Y (x, y) fY (y) = { 1/2 (2−y)/2 = 1 2−y , x ∈ (y, 2) 0, otherwise Note that the conditional PDF of X does not depend on X (is flat), and is actually Uniform(y, 2). Page 9 of 10 (contd...) 10. Consider two RVs X, Y where X has marginal PDF fX(x) = { 2x, x ∈ (0, 1) 0, otherwise and Y has conditional PDF fY |X(y|x) = { 1/x, y ∈ (0, x) 0, otherwise (i.e. Y given X is uniformly distributed in (0, X).) (a) (10 points) Find the marginal PDF of Y ; are X and Y independent? (justify your answer). (b) (10 points) Find the value of the conditional probability P(X > 3 4 |Y = 1 2 ). Solution: (a) The joint PDF of X, Y is given by fX,Y (x, y) = fX(x)fY |X(y|x) = { 2x/x = 2, 0 < y < x < 1 0, otherwise The marginal PDF of Y is found by integrating out X from the joint PDF fY (y) = ∫ ∞ x=−∞ fX,Y (x, y)dx = {∫ 1 x=y 2dx = 2(1− y), y ∈ (0, 1) 0, otherwise The RVs are NOT independent, since the conditional and marginal PDFs of Y are different (moreover, the conditional range of Y depends on X). (b) Note that the joint distribution is Uniform over the triangle 0 < y < x < 1. If you condition on Y = 1/2, then X will range uniformly over (1/2, 1); hence, the conditional probability of X > 3/4 is P (X > 3 4 |Y = 1 2 ) = 1/2. You can verify this more formally using the conditional PDF of X given Y fX|Y (x|y) = fX,Y (x, y) fY (y) = { 2 2(1−y) = 1 1−y , x ∈ (y, 1) 0, otherwise Note that the conditional PDF of X does not depend on X (is flat), and is actually Uniform(y, 1). For Y = 1/2, we have fX|Y (x|1/2) = 11−1/2 = 2,∀x ∈ (1/2, 1). P(X > 3 4 |Y = 1 2 ) = ∫ 1 3/4 fX|Y (x|1/2)dx = ∫ 1 3/4 2dx = 2(1− 3/4) = 1/2 Page 10 of 10 Total marks: 200
欢迎咨询51作业君