Section 3.6 Recall that y –1/2 e –y dy = 0 (Multivariable Calculus is required to prove this!) (1/2) = Perform the following change of variables in the integral: w = 2y y =dy = < w < < y < 0 2 edw= – w 2 / 2 w 2 / 2 0 0 w dw From this, we see that 0 2 e ————— dw = 1 – w 2 / 2
– 2 e ————— dw = 2 – w 2 / 2 – e ——— dw= 1 2 – w 2 / 2 Let – < a < and 0 < b < , and perform the following change of variables in the integral: x = a + bww =dw = < x < < w < A random variable having this p.d.f. is said to have a normal distribution with mean and variance 2, that is, a N( , 2 ) distribution. A random variable Z having a N(0,1) distribution, called a standard normal distribution, has p.d.f. e f(z) =——— for – < z < 2 – z 2 / 2 We shall come back to this derivation later. Right now skip to the following:
We let (z) = P(Z z), the distribution function of Z. Since f(z) is a symmetric function, it is easy to see that (– z) = 1 – (z). Tables Va and Vb in Appendix B of the text display a graph of f(z) and values of (z) and (– z). Important Theorems in the Text: If X is N( , 2 ), then Z = (X – ) / is N(0,1). Theorem If X is N( , 2 ), then V = [(X – ) / ] 2 is 2 (1). Theorem We shall discuss these theorems later. Right now go to Class Exercise #1:
P(Z < 1.25) = (1.25) = P(Z > 0.75) = 1 – (0.75) = P(Z < – 1.25) = (– 1.25) =1 – (1.25) = P(Z > – 0.75) = 1 – (– 0.75) =1 – (1 – (0.75)) = (0.75) = The random variable Z is N(0, 1). Find each of the following: P(– 1 < Z < 2) = (2) – (– 1) = (2) – (1 – (1)) = (– 1) – (– 2) =(1 – (1)) – (1 – (2)) = P(Z < 6) = (6) = practically 1 P(– 2 < Z < – 1) =
a constant c such that P(Z < c) = P(Z < c) = (c) = c = 0.23 a constant c such that P(Z < c) = P(Z < c) = (c) = – (– c) = (– c) = – c = 1.16c = – 1.16 a constant c such that P(Z > c) = 0.25 a constant c such that P(Z > c) = 0.90 P(Z > c) = – (c) = 0.25c 0.67 P(Z > c) = – (c) = 0.90 (– c) = 0.90 – c = 1.28c = – 1.28
1.-continued P(Z > z ) = 1 – (z ) = z 0.10 = z 0.90 = – z 0.10 = – a constant c such that P(|Z| < c) = 0.99 P(– c < Z < c) = 0.99P(Z < c) – P(Z < – c) = 0.99 (c) – (– c) = 0.99 (c) – (1 – (c)) = 0.99 (c) = c = z = z 0.10 z 0.90 P(Z > z ) = 1 – (z ) = (–z ) = 1 – (–z ) = 1 – P(Z > –z ) = 1 – z 1– = –z
– 2 e ————— dw = 2 – w 2 / 2 – e ——— dw= 1 2 – w 2 / 2 Let – < a < and 0 < b < , and perform the following change of variables in the integral: x = a + bww =dw = < x < < w < (x – a) / b – (1/b) dx – e ———— dx = 1 b 2 (x – a) 2 – ——— 2b 2 The function of x being integrated can be the p.d.f. for a random variable X which has all real numbers as its space.
The moment generating function of X is M(t) = E(e tX ) = – e tx e ———— dx = b 2 (x – a) 2 – ——— 2b 2 – e ———— dx = b 2 (x – a) 2 – 2b 2 tx – —————— 2b 2 – exp{} —————————— dx b 2 (x – a) 2 – 2b 2 tx – —————— 2b 2 Let us consider the exponent (x – a) 2 – 2b 2 tx – ——————. 2b 2 (x – a) 2 – 2b 2 tx – —————— = 2b 2 x 2 – 2ax + a 2 – 2b 2 tx – ————————— = 2b 2 x 2 – 2(a + b 2 t)x + (a + b 2 t) 2 – 2ab 2 t – b 4 t 2 – —————————————————= 2b 2
[x – (a + b 2 t)] 2 – 2ab 2 t – b 4 t 2 – ————————————.Therefore, M(t) = 2b 2 – exp{} —————————— dx = b 2 (x – a) 2 – 2b 2 tx – —————— 2b 2 – exp{} —————————— dx = b 2 [x – (a+b 2 t)] 2 – —————— 2b 2 b 2 t 2 exp{at + ——} 2 b 2 t 2 at + —— 2 e b 2 t 2 at + —— 2 efor – < t < M (t) = M (t) = b 2 t 2 at + —— 2 (a + b 2 t) e b 2 t 2 at + —— 2 (a + b 2 t) 2 e+ b 2 t 2 at + —— 2 M(t) = b 2 e
E(X) = M (0) =E(X 2 ) = M (0) = aa 2 + b 2 Var(X) =a 2 + b 2 – a 2 = b 2 Since X has mean = and variance 2 =, we can write the p.d.f of X as ab2b2 e f(x) =———— for – < x < 2 (x – ) 2 – ——— 2 2 A random variable having this p.d.f. is said to have a normal distribution with mean and variance 2, that is, a N( , 2 ) distribution. A random variable Z having a N(0,1) distribution, called a standard normal distribution, has p.d.f. e f(z) =——— for – < z < 2 – z 2 / 2
We let (z) = P(Z z), the distribution function of Z. Since f(z) is a symmetric function, it is easy to see that (– z) = 1 – (z). Tables Va and Vb in Appendix B of the text display a graph of f(z) and values of (z) and (– z). Important Theorems in the Text: If X is N( , 2 ), then Z = (X – ) / is N(0,1). Theorem If X is N( , 2 ), then V = [(X – ) / ] 2 is 2 (1). Theorem 3.6-2
2.The random variable X is N(10, 9). Use Theorem to find each of the following: P(6 < X < 12) = 6 – 10 X – – 10 P( ——— < ——— < ———— ) = P(– 1.33 < Z < 0.67) = (0.67) – (– 1.33) = (0.67) – (1 – (1.33)) = – (1 – ) = P(X > 25) = X – – 10 P( ——— > ———— ) = 3 3 P(Z > 5) = 1 – (5) = practically 0
P(|X – 10| < c) = 0.95 X – 10 c P( ——— < — ) = P(|Z| < c/3) = 0.95 (c/3) – (– c/3) = 0.95 (c/3) – (1 – (c/3)) = 0.95 (c/3) = c/3 = z = 1.960c = continued a constant c such that P(|X – 10| < c) = 0.95
3.The random variable X is N(–7, 100). Find each of the following: P(X > 0) = X P( ——— > —— ) = P(Z > 0.7) = 1 – (0.7) = a constant c such that P(X > c) = 0.98 P(X > c) = 0.98 X + 7 c +7 P( —— > —— ) = P(Z > (c+7) / 10) = – ((c+7) /10) = 0.98 ((c+7) /10) = 0.02 (c+7) /10 = z 0.98 = – z 0.02 = – c = – 27.54
the distribution for the random variable Q = X X + 49——— 100 From Theorem 3.6-2, we know that Q = X X + 49 —————— = 100 X + 7 —— 10 2 must have a distribution. 2 (1) 3.-continued