STA331 TEST 1 – 29 February 2012 Time: 1 hour Marks: 35 Question 1 (7) (i) Define (p) . (ii) 1 (ii) Show that ( ) 2 (iii) Define μ r and μ r for a continuous random variable X . π. Question 2 (18) If the probability density function (p.d.f.) of X is f x ( x ) e x , x 0 0 otherwise (i) Show that M x ( t ) (1 t ) 1 (ii) Hence, obtain μ′ r for this distribution Using the transformation theorem show that the probability density function ( p.d.f .) of y x 2 (a monotonic function over the range x 0) and x having p.d.f. in (i)) is: f y ( y) 1 .e y , 0 y 2 y 0 otherwise (iii) Hence show that E[Y] 2E[X] Question 3 (10 marks) A random variable X has a chi-square distribution with n degrees of freedom with p.d.f . f (x) 1 n 2 2 ( n2 ) n 1 x2 e 0 x 2 0x otherwise n Show that the cumulant generating function (c.g.f.) of X is K X ( t ) n (1 2t ) 2 Hence, show that for the above chi-square distribution μ1 n ; μ 2 var( X) 2n Hint: n (1 x) x x 2 x3 2 3 for x 1 Memorandum Question 1 [7] (i) Define (p) . 1 (ii) Show that ( ) 2 π. (ii) Define μ r and μ r for a continuous random variable X . Solution: (i) (p) x p1e x dx for p 0 0 1 2 (1) 1 2 (ii) ! π 1 1 x 2 e x dx 2 0 put x = 1 2 2 2y e y dy y 0 1 2 2 2y dy e 2 0 1 2 y then dx = y dy 2 2 1 y2 e 2 dy 0 and as e 1 y2 2 is an even function 2 . 2π π 2 (4) (iii) If X has pdf f x (x) μ r x r f x (x)dx ; μ r (x μ1 ) r f x (x)dx x x (2) Question 2 (18) If the probability density function (p.d.f.) of X is f x ( x ) e x , x 0 0 otherwise (i) Show that M x ( t ) (1 t ) 1 (ii) Hence, obtain μ′ r for this distribution Using the transformation theorem show that the probability density function ( p.d.f .) of y x 2 (a monotonic function over the range x 0) and x having p.d.f. in (i)) is: f y ( y) 1 2 y .e y , 0 y 0 otherwise (iii) Hence show that E[Y] 2E[X] Solution: 0 0 (i) M x ( t ) E[e xt ] e xt e x dx e x (1 t ) dx 0 (ii) Now e x (1t ) ] If t 1 (1 t ) 0 1 (1 t ) 1 (1 t ) 1 1 t t 2 ... t r ... (1 t ) t 1 1 t 2! Coefficient of Alternatively dt 2 d3 dt 3 dr dt μ r t2 t2 ... r! ... 2! r! tr μr r! r 1, 2, ... r! d d μ r (t ) (1 t ) 1 1(1 t ) 2 (1) 1(1 t ) 2 dt dt d2 In general (3) r (1 t ) 1 2(1 t ) 3 (1) 1.2(1 t ) 3 (1 t ) 1 1.2(3)(1 t ) 4 (1) 1.2.3(1 t ) 4 (1 t ) 1 1.2. ... .r (1 t ) (r 1) (1) dr M x ( x ) r!(1 t ) ( r 1) r! r r dt t 0 dt t 0 dr (iii) For Y 0 Fr ( y) P[Y y] P[ x 2 y] 0 (4) f r ( y) dFy ( y) dy 0 ,y 0 For Y 0 pdf of Y is f y ( y) f x ( x ) dx dy now Y x 2 x Y x 0 1 e x 2 y dx 1 dy 2 y 0 =e y . 1 y>0 2 y y0 (4) (iv) E(X) xf x ( x )dx 0 0 0 xe x dx x 21e x dx (2) 1! 1 and E(Y) yf y ( y)dy y 0 Now put Z y 0 y z2 1 e y dy 2 y and dy 2zdz E( Y) z 2 . 0 1 z .e 2zdz z 31e z dz (3) 2! 2 2z 0 E(Y) 2E(X) 2.1 2 0 0 Alternatively E(Y) E(X 2 ) x 2 f x ( x )dx x 2 e x dx x 31e x dx (3) 2! 2 0 (7) Question 3 (10 marks) A random variable X has a chi-square distribution with n degrees of freedom with p.d.f . f (x) 1 n 2 2 ( n2 ) 0 n 1 x2 e x 2 otherwise 0x (i) Show that the cumulant generating function (c.g.f.) of X is n K X ( t ) n (1 2t ) 2 Hence, show that for the above chi-square distribution μ1 n ; μ 2 var( X) 2n (ii) x 2 x3 2 3 Hint: n (1 x) x for x 1 Solution: (i) M X ( t ) E[e xt ] e xt f X ( x )dx e xt 0 0 2 n2 ( n2 ) 0 dx 1 2 n 1 2 n2 ( n2 ) 0 1 ( 2 t) 2 1 n dx t y2 2 dx dy n 1 1 [2( 12 t )] 2 i) 2 n2 ( n2 ) x2 e n 1 ( 1 t ) x x2 e 2 1 1 Put y ( t ) x 2 M X (t) n 1 x 1 e yt dy ( 12 t ) 1 n ( n2 ) n 2 2 ( n2 ) ( 12 t ) 2 1 n (1 2t ) 2 (5) (Use K or C for cumulant referral) n K X (t ) n M X (t ) n(1 2t ) 2 (2) n ii) K X (t ) n M X (t ) n(1 2t ) 2 n (2t ) 2 (2t ) 3 (2t ) 2 2 3 nt 2n t 1 2 t2 2 Coefficient of t in expression of K x ( t ) : μ1 n Coefficient of t2 in expression of K x ( t ) : μ 2 2n 2! (3)