Documente Academic
Documente Profesional
Documente Cultură
Pilani Campus
Chapter 4
Continuous Distribution
Continuous Random Variables
Definition: A random variable is continuous
if it can assume any real numbers and some
interval (or intervals) of real numbers and
the probability that it assume any specific
value is 0 (zero).
for any a b.
BITS Pilani, Pilani Campus
Note :
f ( x)dx k 1,0 k
=
f (t ) dt
4. lim F ( x)
x
x
lim
x
f (t )dt
f (t )dt 1
f ( x)dx 1
1/(c-a)
x
X=a X=c
BITS Pilani, Pilani Campus
( ii ) Shade the area in the graph of part (b)
that represents P[X (a + c)/2].
f(x)
(a+c)/2 X= c x
X= a
(c ) Find the probability pictured in part: ii
a c
a c 2
1
P X
2
a
ca
dt 0.5
xf ( x ) dx
| H ( x ) | f ( x ) dx
is finite.
BITS Pilani, Pilani Campus
…continue
3. Var X = E[X2] – (E[X])2 = 2 , where
E[ X ] x f ( x)dx
2 2
Variance is shape parameter in the
sense that a random variable with small
variance will have a compact density;
one with a large variance will have a
density that is rather spread out or flat.
BITS Pilani, Pilani Campus
…continue
= E[ etX] = mX (t) = tx
e f ( x ) dx
c
1
E[e ] e
tX tx
dx
a
c a
1 e e
tc ta
, t 0
t ca
BITS Pilani, Pilani Campus
Use definition to find mean and
variance can be found.
ac
E( X )
2
(c a ) 2
Var ( X )
12
BITS Pilani, Pilani Campus
Section 4.1, page no 139, 6
1 a
f (x) ,
p a ( x b)
2 2
x , b , a 0
is said to have a Cauchy distribution with
parameters a and b. This distribution is
interesting in that it provides an example of a
continuous random variable whose mean does
not exist. Let a = 1, b = 0 to obtain a special
BITS Pilani, Pilani Campus
Cauchy distribution with density
1 1
f ( x) , x
p 1 x 2
Show that | x | f ( x) dx does not exist.
Sol:
1 1
| x | f ( x ) dx p 1 x 2
| x | dx
=0, for x 0,
is said to have a Gamma Distribution with
parameters a and b,for a > 0,b > 0.
0
is called the Gamma function.
Properties :
1. (1) = 1
2. (a) = (a -1) (a -1 ); where a>1 is real
number.
3. (n) = (n-1)! (Factorial of n-1), n = 1,2,3….
4. (1/2)=p
BITS Pilani, Pilani Campus
1. (1) = 1
Proof: By definition of Gamma function, we
have
(1) e z dz; z 0, a 1,
z 0
0
e dz 1
z
(a 1)(a 1)
3. (a) (a 1)!
Since,
(a) (a 1)(a 1) ...
(a 1)(a 2)...(1) (a 1)!
BITS Pilani, Pilani Campus
1
1
z e z dz p
2
2 0
1
2 e dt
t 2
2 0
1
x e , x 0,a 0, b 0
a 1 x / b
f ( x) (a ) b a
0,
other wise
x
Let t dx b dt , and x b t
b
1
a
a 1 a 1 t
b t e b dt
(a ) b 0
BITS Pilani, Pilani Campus
a
b
a
a 1 z
f ( x ) dx z e 1
(a ) b 0
Hence f ( x) is a pdf
(
1
t)x
1
x
a 1 b
e dx
(a ) b a
0
x zb
let z (1 bt ) x
b (1 b t )
bdz
and dx
(1 bt )
BITS Pilani, Pilani Campus
1 bz a 1 z bdz
a
( ) e
(a ) b 0 1 b t (1 bt )
a
1 b
a
a 1 z
a
z e dz
(a) b (1 bt ) 0
dt t 0
a(a 1)b 2
Var ( X ) E[ X ] E[ X ]
2 2
0.1
0.05
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
x
BITS Pilani, Pilani Campus
For a > 1, the maximum value of the
density occurs at the point x = (a – 1)b.
0.1
0.05
0
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20
x
BITS Pilani, Pilani Campus
Exercise 4.3 (Page No: 143)
Let X be a gamma random variable with
a = 3 and b = 4
(i) What is the expression for the density
for X x x
1 a 1 b 1 2 4
f ( x) a
x e x e , x 0
(a ) b 2!4 3
x
1 2
x e , x0
4
128
BITS Pilani, Pilani Campus
(ii) What is the moment generating
function for X
a 3 1
m X (t ) (1 bt ) (1 4t ) , t
4
(iii) Find mean, variance and standard
deviation
ab 12
ab 48
2 2
48 6.9282
BITS Pilani, Pilani Campus
Exponential Distribution
In Gamma Distribution, put a = 1, we
get
1 x b
e , x 0, b 0
f ( x) b
0 ,
elsewhere
s x x
b b
e 1 e x0
0
BITS Pilani, Pilani Campus
The c.d.f. of exponential
distribution
0, x0
F ( x)
x
1 e
b
x0
F (w) f ( w) e w
f ( x) (a ) b a
0,
other wise
f ( x) (1 / 2) 2
f(x)
1/ 2
0,
other wise
0.05
-1.94E-16
0 5 10 15 20 25
x
BITS Pilani, Pilani Campus
We do not have explicit formula for CDF
F of X2 . Instead values are tabulated on
page no. 695-696 as below (F occurs in
margin here, and related value of r.v.
inside the table):
≥
2
2
P[X a] =α
Probability to the right of
2
a is α.
for exampleP[X210 ≥ 2
0.05 ] = 0.05
BITS Pilani, Pilani Campus
Probability to the right of is α.
2
a
0.1
f(x)
0.05
0.05
0
0 5 10 15 20 25
P[X210 ≥
2
0.05 ] = 0.05 x
= 1 - P[X210 ≥
2
0.95 ] = 0.95
BITS Pilani, Pilani Campus
P[X2 ≤ t]
F 0.90 0.95 0.975
f(x)
0.05
0.05
-1.94E-16
0 5 10 15 20 25
For 10 degrees of freedom, 2
.05 18.3 x
P[X210 ≥
2
0.05 ] = 0.05
BITS Pilani, Pilani Campus
38. Consider a chi – squared random
variable with 15 degrees of freedom.
2
(i) What is the mean of X 15 ?
Sol: Chi square distribution is Gamma
with a = /2 & b = 2, Mean = ab = =15
and 2 = ab2 = 30
BITS Pilani, Pilani Campus
(ii) What is the expression for the density for
2
X ?
15
b 2, a / 2, hence
1 (15 / 2 ) 1 x / 2
f ( x) x e ,x 0
(15 / 2) 215 / 2
0 otherwise
Sol:
a
mX (t ) (1 bt ) (1 2t ) , t 1 / 2
15 / 2
P[X2 < t]
F 0.005 0.010 .025 0.900
2
P[ X 15 22.3] 1 P[ X 15
2
22.3]
1 0.900 0.10
BITS Pilani, Pilani Campus
(v) Find P[6.26 X 22.3]
2
15
P[X2 < t]
F 0.005 0.010 .025 0.900
P[6.26 X 15 22.3]
2
P[X 2 2
> r] = r.
15
P[X215 > 20.01] = 0.01.
P[X215 ≤ 20.01] = 0.99.
0.01 30.6
2
BITS Pilani, Pilani Campus
4.4 The Normal Distribution
A random variable X with density f(x)
is said to have normal distribution
with parameters and > 0, where
f(x) is given by:
x 2
1 2 2
f ( x) e , x, (, ); 0.
2p
BITS Pilani, Pilani Campus
(i) f ( x) 0 x (-, )
(ii) f(x)dx 1 2
- 1 x
1
e
2
To prove dx 1
2p - 2
x
let z dx dz
2
1 x 1
1 1 z2
e
2
dx e 2
dz
2p 2
- 2p -
BITS Pilani, Pilani Campus
1 1
1 z2 1 z2
e 2
dz 2 e 2
dz
2p - 2p 0
1 1
z2 z2
e dz e dz I
0
2
0
2
1 1
x2 y2
I I e 2
dx e 2
dy
0 0
1
( x2 y2 )
e 2
dxdy
0 0 BITS Pilani, Pilani Campus
1
( x2 y2 )
e dxdy
2
0 0
lim p /2 R 1 2
(r )
q
e rdrd
2
R
0 0
p /2 R /2
2
lim e w
dw dq
R
0 0
p
2 BITS Pilani, Pilani Campus
p
i.e. I
2
1 1
1 z2 1 z2
e 2
dz 2 e 2
dz
2p - 2p 0
2 p
1
p 2
BITS Pilani, Pilani Campus
Mean and Standard deviation
for Normal distribution
Theorem: Let X be a normal
random variable with parameters
and . Then is the mean of X and
is its standard deviation.
Proof deferred.
z2
1
mZ (t ) E[e ]
Zt zt 2
e e dz
2p
z2
1 zt
2
e dz
2p
Let z - t = w dz = dw
w2
t2
t2
1
mZ (t ) e 2
2p e
2
dw e 2
t 2 2
t
e 2
BITS Pilani, Pilani Campus
Mean and Variance for Normal r.v.
d
Mean E[ X ] m X (t )
dt t 0
2
d
and E[X ] 2 m X (t )
2 2 2
dt t 0
2
BITS Pilani, Pilani Campus
(i) If X is N(, )
c a
P(a X c) FZ FZ
(ii) FZ ( z) 1 FZ ( z)
(iii ) FZ ( z ) 0 if z -3.5
FZ ( z ) 1 if z 3.5
BITS Pilani, Pilani Campus
BITS Pilani, Pilani Campus
Symmetry
F(2μ-a)=1-F(a).
a-μ
2μ-a μ a
(vi) P(Z ≥ zr ) = r
P(Z ≥ z.10 ) = 0.10
P(Z < z.10 ) = 0.90
z.10 = 1.28
X - 10 12
P(X 10) P
4
P(Z -0.5) 1 - P(Z 0.5)
1 - F(-0.5) 1 0.3085
a b
FZ FZ .5 0.25
a
0.67( why)
a (4)( 0.67) 12 12 2.68 9.32
BITS Pilani, Pilani Campus
Section 4.4, page no 145, 41
Most galaxies take the form of a flatten disc,
with the major part of the light coming from this
very thin fundamental plane. The degree of
flattening differs from galaxy to galaxy. In the
Milky Way Galaxy most gases are concentrated
near the center of the fundamental plane. Let X
denote the perpendicular distance from this
center to a gaseous mass. X is normally
distributed with mean 0 and standard deviation
100 parsecs. (A parsec is equal to
approximately 19.2 trillion miles.)
BITS Pilani, Pilani Campus
(a) Sketch a graph of the density for X. Indicate on
this graph the probability that a gaseous mass
is located within 200 parsecs of the center of
the fundamental plane. Find this probability.
(b) Approximately what percentage of the gaseous
masses are located more than 250 parsecs
from the center of the plane?
(c) What distance has the property that 20% of the
gaseous masses are at least this far from the
fundamental plane?
(d) What is the moment generating function for X?
BITS Pilani, Pilani Campus
Log-Normal Distribution
P(3 X 8)
P(3 0.5 X 8 0.5)
(correction to continuity )
2.5 6 X np 8 .5 6
P
20 .3 .7
npq 20 .3 .7
P(1.71 Z 1.22)
FZ (1.22) FZ (1.71) 0.8888 0.0436
BITS Pilani, Pilani Campus
53. Although errors are likely when taking
measurements from photographic images, these
errors are often very small. For sharp images with
negligible distortion, errors in measuring
distances are often no larger then 0.0004 inch.
Assume that the probability of a serious
measurement error is 0.05. A series of 150
independent measurements are made. Let X
denote the number of serious errors made.
f ( x)
e s
s
x
, x 0,1,2,3,.......
x!
Then for large values of s, X is
approximately normal with mean s and
variance s.
100
500 k 100, i.e., k
500
10000
k
2
20
500
1 1 19
Therefore, 1 2 1
k 20 20
P900 X 1100
19
20
BITS Pilani, Pilani Campus
X = number of bulletproof vests out of 405
which fail to meet puncture standards.