Sie sind auf Seite 1von 61

STATISTIC

1.1 Sample space, S

The set of all possible outcomes of a random experiment is called the sample
space. It is represented as S.
Each outcome in the sample space is called an element / member. Thus, the sample space
s of possible outcome when a coin is tossed is written as
S

Random experiment element / sample

S = {H ,T } H-head
T-tail

Set of possible outcomes, S

Random experiment – an experiment that can result in different outcomes, even though it is
repeated in the same manner.

Example 1

Consider a random experiment of tossing a die. If we are interested in the numbert that shows on
the top face, the sample space,

S1 = {1,2,3,4,5,6,}

If we are interested only in whether the number is odd or even, the sample space

S 2 = { odd , even }

*more than one sample space can be used to describe the outcomes of an experiment.

In some experiments, it is helpful to list the elements by using a tree diagram.

Example 2

Suppost that tree items are selected at random from a manufacturing process. Each item is
inspected and classified defective, D or nondefective, N

1
1st item 2nd item 3rd item outcomes

DDD
D

D
N DDN

D
D DND
N

N N
DNN

D
NDD

N
NDN

N D
NND

N
NNN

The sample space, S = { DDD , DDN , NNN }

1.2 Events

2
An event is a subset of a sample space

From example 2, we may be interested in the event B that the number of defectives is greater than
1. The event B is written as

B = { DDN , DND , NDD , DDD } S B



 

This subset represents all the elements for which the event is true.

Example

Throwing a die: S = {1,2,3,4,5,6}


Event, E1 = (odd number)= {1,3,5}
Event, E2 = (prime number)= {2,3,5}

The complement of an event A with respect to S is the subset of all elements


of S , that are not in A . It is stated by a symbol S ′
Example 3

Consider a sample space, S = { pizza , burger , lasagna }

Let A = { pizza }
The complement of A is A′ = {burger , lasagna }
The intersection of two events A and B , stated by the symbol A ∩ B , is the event containing
all elements that are common to A and B .

Let S = {1,2,3,4,5,6} , A = {2,4,6} , B = {4,5,6}


A ∩ B = { 4,6}

S A B

2 4
 1 6 5 3

Two events A and B are mutually exclusive or disjoint if A ∩ B = 0/ , that is, if A and

3
B have no elements in common.

Let M ={a, e, i, o, u} and N = {r , s, t}

Then M ∩ N = 0/ = { }

The union of the two events A and B , stated by the symbol A ∪ B , is the event
containing all the elements that belong to A or B or both.

Let A = {a, b, c} and B ={b, c, d , e}

Then A ∪B = {a, b, c, d , e}

1.3 Probability of an event

The probability of event A is

n ( A)
P ( A) = n( A) - Number of elements of event
n( S )
A
n( S ) - Number of elements of sample
space, S

Example 4

A statistic class for engineers consist of 10 mechatronic, 8 electrical and 12 electronic


engineering students. If a student is randomly selected, find the probability that the
student is

a) an electrical engineering major


b) an electrical or electronic major

Solution

Denote A,B,C as the student majoring mecha, electrical and electronic major

n( B ) 8
a) P ( B ) = =
n( S ) 30

b) P( B ∪ C ) = n( B ∪ C ) = 20 = 2
n( S ) 30 3

4
Adaptive rule
A∩ B

If A and B are two events, then S

P ( A ∪ B ) = P ( A) + P ( B ) − P ( A ∩ B )

A B

Corallary

If A and B are mutually exclusive, A∩B = 0


then P( A ∪ B ) = P( A) + P( B ) P ( A ∩ B ) = P ( 0/ )
=0

1.4 Conditional Probability

The probability of an event B occuring when it is known that some event A has
occurred is called a conditional probability and it is stated by P(B A)

P (B A) is usually read ̀the probability that B occurs given that A occurs’ or the
probability B given A’

The conditional probability of B , given A , stated by is defined by


P (B A)
Is defined by

P( A ∩ B )
P ( B A) = provided P( A) > 0
P ( A)

Example 5

The probability that a regularly scheduled flight departs on time is P( D ) = 0.83 ; the
probability that it arrives on time is P( A) = 0.82 ; the probability that is departs and
arrives on time is P( D ∩ A) = 0.78 . Find the probability that a plane

5
a) arrives on time that it departed on time
b) departed on time given that it has arrived on time

solution

P ( D ∩ A) 0.78
a) P ( A D ) = = = 0.94
P( D ) 0.83
b) P( D A) = P( D ∩ A) = 0.78 = 0.95
P ( A) 0.82

Example 6

Sample space S is the population of adults in a small town and categorized according to
gender and employment status. (table 1)

Employed, E Unemployed, E ′ Total


Male,F 460 40 500
Female,F 140 260 400
Total 600 30 900

Find,
The probability of a men is chosen given that the one chosen is employed

P( M E ) =
460 23
=
600 30
Or

P( M ∩ E ) n( M ∩ E ) 460
P( M E ) = , P( M ∩ E ) = =
P( E ) n( s ) 900

n( E ) 600
P( E ) = =
n( s ) 900
hence

P( M ∩ E )
P( M E ) =
P( E )

460
23 as before
= 900 =
600 30
900

1.5 Multiplicative Rule

6
If in an experiment the events A and B can both occur, then

P ( A ∩B ) = P ( A ) P ( B A) , provided P ( A) > 0

Two events A and B are independent if and only if

P ( A ∩ B ) = P ( A ) P ( B ), , P ( B A ) =P ( B )
(independent)

Example 7

In the card tossing experiment, 2 cards are drawn in succession ( 5 yellow and 10 blue
cards) without replacement.

1st card 2nd card Outcomes probability

7
4
P ( y21 y1 ) =
14
yellow y2 y1 y2 P( y1 ∩ y2 )

blue b2 y1b2 P( y1 ∩ b2 )

10
P( b21 y1 ) =
14

5
P ( y1 ) =
15

yellow y1

blue b1

10
y ( b1 ) =
15

5
P ( y21b1 ) =
14
yellow y2 b1 y2 P( b1 ∩ y2 )

b2 b1b2 P( b1 ∩ b2 )
Blue
9
P ( b21b1 ) =
14

5 4
y1 y2 : P ( y1 ∩ y2 ) = P ( y1 ) × P ( y21 y1 ) = × =
15 14

5 10
y1b2 : P( y1 ∩b2 ) = P ( y1 ) × P ( b21 y1 ) = × =
15 14

8
10 5
b1 y2 : P( b1 ∩ y2 ) = P( b1 ) × P( y21b1 ) = × =
15 14

10 9
b1b2 : P ( b1 ∩b2 ) = P ( b1 ) × P( b21b1 ) = × =
15 14

With replacement

1st card 2nd card outcomes probability

5
P( y21 y1 ) = y1 y2 P( y1 ∩ y2 )
15
yellow y2

Blue b2 y1b2 P( y1 ∩ b2 )
10
P( b21 y1 ) =
15
5
P ( y1 ) =
15
yellow y1

Blue b1
10
y ( b1 ) =
15
5
P ( y21b1 ) = b1 y2 P( b1 ∩ y2 )
15
yellow y2

blue b2 b1b2 P( b1 ∩ b2 )
10
P ( b21b1 ) =
15

5 5
y1 y2 : P ( y1 ∩ y2 ) = P( y1 ) × P ( y2 ) = × =
15 15

5 10
y1b2 : P ( y1 ∩ b2 ) = P ( y1 ) × P ( b2 ) = × =
15 15

9
10 5
b1 y2 : P ( b1 ∩ y2 ) = P ( b1 ) × P ( y2 ) = × =
15 15

10 10
b1b2 : P ( b1 ∩b2 ) = P ( b1 ) × P ( b2 ) = × =
15 15

Since, the first card is replaced, hence

5 5
P ( y21 y1 ) = and P ( y2 ) =
15 15

That is P ( y21y1 ) = P ( y2 ) When this is true, the events y1 and


y2 are said to be independent

And this is also true for:

P( b21 y1 ) = P( b2 ) , P ( y21b1 ) = P ( y2 ) , P( b21b1 ) = P( b2 )

1.6 Bayes Rule

Let us return to table 1 (example 6)


s 900
Event E – adult employed, n( E ) = 600 E
Event E ′ – adult unemployed, n( E ′) = 300 E’
E

Suppose that we are now given the additional information that 36 of those employed and
12 of those unemployed are members of swimming club.

Event A-members of swimming club


S
E E∩A
n( A E ) =36
E’ E ′ ∩ A n( A E ′) = 12
Venn diagram for event A, E , E ′
We wish to find the probability of adult selected as a member of
the swimming club, P ( A ) = P ( E ∩ A) + P ( E ′ ∩ A) 600 2
P( E ) = =
900 3
From conditional probability, P( E ∩ A) = P( E ) P ( A E )
P( E ′ ∩ A) = P ( E ′) P ( A E ′) P( E ′) =
300 1
=
900 3
P( A E ) =
36 3
=
600 50
110
P ( A E ′) =
12
=
300 25
Hence ,

P ( A) = P ( E ) P ( A E ) + P ( E ′) P ( A E ′)

 2 3  1 1 
= × + × 
 3 50   3 25 

4
=
75
P(
A E )
P ( E ∩A) = P ( E ) P ( A E )
A
A′ P ( E ∩ A′)

P (E )
E

E′ A′
P( E ′)
P ( A E ′)
P ( E ′ ∩ A) = P ( E ′) P ( A E ′) A

P( E ′ ∩ A′)
P ( A′) = 1 − P ( A)

The generalization of this illustration (where the sample space is partitioned into K
subsets (2 subsets in this case) is covered by

‘the theorem of total probability’


or ‘rule of elimination’

If the events B1 , B2 , Bk constitute a partition of the sample space S such that

P( Bi ) =/ 0 for i =1,2, k , then for any event A of S ,

11
k k
P( A) = ∑ P( Bi ∩ A) = ∑ P( Bi )P ( A Bi )
i =1 i =1

Theorem of total probability

Example 8

In a certain assembly plant, three machine B1 , B2 and B3 make 30%,45% and 25%,
respectively, of the products. It is known that 2%, 3% and 2% of the products made by
each machine, respectively are defective. What is the probability that the product selected
is defective.

Solution
S

B1 D B2
B1 ∩ D B2 ∩ D

B3 ∩ D
B3

P ( D1B1 ) = 0.02
B1 DP( B1 ∩ D )
P( B1 ) = 0.3
P( B2 ) = 0.45 D′ P( B1 ∩ D′ )
P ( D1B2 ) = 0.03
B2
DP( B2 ∩ D )

B3
P( B3 ) = 0.25
P( D1B3 ) = 0.02
D′ P( B2 ∩ D′ )

D P( B ∩ D ) 3

D′
P( D ) = P ( B ) P ( D1B ) + P ( BP)(PB(3D∩1B
1 1 2D′) + P( B ) P( D1B )
2 3 3
= ( 0.03 )( 0.02 ) + ( 0.45 )( 0.03 ) + ( 0.25 )( 0.02 )
= 0.0245

P( D′) = 1 − P( D )

Rule of elimination (total probability) to find P( A) or P( A′)


( P ( A Bi ) provided)

12
Now, we wish to find the conditional probability,

P ( Bi A)

To answer this, we must know Bayes’ rule.

Bayes rule: If the events B1 , B2 , Bk constitute a partition of the sample space s such
that P( Bi ) =/ 0 for i = 1,2 k , , then for any event A in S such that P( A) =/ 0

P( Br ∩ A) P ( Br ) P ( A Br )
P( Br A) = k
= k

∑ P ( B ∩ A) ∑ P ( B ) P ( A B )
i =1
i
i =1
i i

r =1,2, k

Example 8 (continue)

If the product was chosen randomly and found to be defective, what is the probability that
it was made by machine B3 ?

Solution

P( B3 ) P( D1B3 )
P( B31D ) =
P( B1 ∩ D ) + P( B2 ∩ D ) + P( B3 ∩ D )

P( B3 ) P( D1B3 )
=
P( B1 ) P( D1B1 ) + P( B2 ) P( D1B2 ) + P( B3 ) P( D1B3 )

P ( B3 ) P( D1B3 )
=
P( D )
=
( 0.25 )( 0.02 )
0.0245

2.1 Concept of random variable

It is important to allocate a numerical description to the outcome. For example, when two
electronic components are tested

Component 1 component 2 Outcomes

13
D DD

D N DN D-defective
N-Non defective
N
D ND

N NN

The sample space, s = { DD , DN , ND , NN }

Let say, we concern with the number of defectives that occur

Sample space DD DN ND NN
x x =2 x =1 x =1 x =0

Here, x is called random variable which associates the number of the defective items.
X Capital letter, X -denote a random variable
Small letter, x -one of its values

A random variable is a function that assigns a real number to each outcome in the
sample space of a random experiment
(count data)
A discrete random variable is a random variable with a finite range or countably infinite.

A continuous random variable is a random variable with an interval (either finite or


infinite) of real numbers of its range.

Example-Continous random variable

1. Y : the amount of milk produces daily by…


0 ≤ y <10 l 10 l ≤ y < 20 l y ≥ 20 l

2. T : the temperature of certain city in…


− 10  c ≤ t < 0 c 0 c ≤ t < 10  c 10  c ≤ t < 20  c

2.2 Probability Distribution


2.2.1 Discrete Probability Distribution

A discrete random variable assume each of its values with a certain probability.

For example, in the case of two electric component are tested,

14
S = { DD , DN , ND , NN }
X : Discrete random variable which represent the number of defectives

X =x 2 1 0
P( X = x ) 1 2 1 1
P( x = 2) = P( X = 1) = = P( X = 0) =
4 4 2 4

It is convenient to represent all the probabilities of a random variable X by a formula:


X = x function of x , f ( x)

For example: f ( x ) = P( X = x ) that is f ( 2 ) = P( X = 2 )

The set of ordered pairs x, f ( x ) is called the probability function or


probability distribution of the discrete random variable X .The function f ( x )
satisfying
1. f ( x ) ≥ 0 2. ∑ f ( x ) = 1 3 P( X = x ) = f ( x )
x

2.2.2 Continuos Probability Distribution

Indealing with continuos variables, f ( x ) is usually called the probability density


function or density function of a continuos random variable of X

The function f ( x ) satisfying


1. f ( x ) ≥ 0 for all x ∈ R

2.
∫ f ( x )dx =1
−∞

3. P( a < x < b ) = ∫a f ( x ) dx =area under f ( x ) between


b
a and b

2.2.3 Mean, variance and standard deviation

Expectation or mean or expected value of a random variable X , represented by E ( x ) or


u is defined as

15
u = E( X ) = ∑x f ( x )
i
i i , if X is discrete

x f ( x )dx , if X is

∫ −∞

continuos

Variance characterizes the variability in the distributions. Variance of a random variable,


X

[ ]
σ 2 = E ( X − u ) = ∑( x − u ) f ( x )
2 2

E( X 2 ) − u2
If X is discrete.

( )
E X 2 = ∑ xi2 f ( xi )
i

[
σ 2 = E ( X −u )2 = ∫ ] ∞

−∞
( x − u ) 2 f ( x ) dx
( )
= E X 2 −u2

If X is continuos
( )
E X 2 = ∫ xi2 f ( x )dx

−∞

Standard deviation (S.D) represented by σ , is the positive square root of variance

σ = + σ2

Question 1

The distribution of the number of imperfections per 10 meters of synthetic fabric is given
by
y 0 1 2
f ( y) 1 3 1
i) Find the expected 5 5 5
number of imperfections, E ( X ) = u , variance, σ and standard deviation,
2
σ
ii) Find, P ( y ≤1) , P( y >1) , P ( y = 0 ) and P( 0 ≤ y ≤ 2) .

Question 2

16
The porportion of impurities in a batch and the density function is given by

f ( x) = x, 0 < x < 1
2 − x ,1 ≤ x < 2
0 , elsewhere

i) Find u ,σ2 and σ


( )
ii) Find P 1 < x < 3 and P x < 3
4 2 2
( )
Solution

2
i) u = E ( y ) = ∑ yi P( yi )
i =0

= y0 P( y0 ) + y1 P( yi ) + y 2 P ( y2 )

( ) ( ) ( )
= 0 1 +13 + 2 1 =1
5 5 5
σ 2 = E( y2 ) − E 2 ( y)
( )
E y 2 = y02 p ( y0 ) + y12 p ( y1 ) + y22 p ( y2 )

5
( ) ( ) ( )
= 0 1 + 1 3 + 22 1
5 5
=7
5

E 2 ( y ) = u 2 =1

∴σ 2 = 7 − 1 = 2/5 σ = 2
5 5
ii) P ( y ≤ 1) = P ( y = 0 ) + P( y = 1)
=1 +3 =4
5 5 5
P( y > 1) = P( y = 2) = 1
5
P( 0 ≤ y ≤ 2 ) = P( y = 0 ) + P ( y = 1) + P ( y = 2 )

= 1 + 3 + 1 =1
5 5 5
P( y = 0 ) = 1
5
2) i) u = E ( x ) = ∞ xf ( x )dx
∫−∞
= ∫ x.xdx + ∫ x ( 2 − x )dx
1 2

0 1

=1

17
σ 2 = E( x2 ) − E 2 ( x)
( )
E x 2 = ∫ x 2 f ( x )dx
−∞

1 2
= ∫ x 2 xdx + ∫ x 2 ( 2 − x ) dx
0 1

= 14
12

∴ σ 2 = 14 − (1) = 1
2
σ = 0.408
12 6

ii)
4
(
P 1 < x< 3
2
)
1 3 27
= ∫1 xdx + ∫ 2 2 − xdx =
4
1 32

P x< 3( 2
)
1 3 7
= ∫ xdx + ∫ 2 2 − xdx =
0 1 8

2.3 Some Discrete Probability Distribution


2.3.1 Discrete Uniform Distribution

The simplest of all discrete probability distribution is one where the random variable
assumes each of its values with an equal probability.

If the random variable X assumes the value x1 , x2  xk with equal probabilities, then
the discrete uniform distribution is given by

f ( x; k ) = 1 , x = x1 , x2 , xk
k
We use the notation f ( x; k ) instead of f ( x ) to indicate that the uniform distribution
depends on the parameter k

Example

18
When a fair die is tossed, each element of the sample space s = {1,2,3,4,5, 6} occurs with
probability 1
6

2.3.2 Binomial Distribution

The Bernoulli Process

The Bernoulli process must possess the following properties:

1. The experiment consist of n repeated trials


2. Each trial results in an outcome that may be classified as a success or a failure.
3. The probability of success, stated by p , remains constant from trial to trial.
4. The repeated trial are independent.

Example 1

The average zinc concentration recovered from a sample of zinc measurements in 36


different locations is found to be 2.6 grams per mililiter. Find the 95% and 99%
confidence intervals fior the mean zinc concentration in the river. Assume that the
population standard deviation is 0.3.

Solution

The point estimate of u is X = 2.6

1)95% C.I: α = 0.025 Z 0.025 =1.96


2
α = 0.05
(Leaving an area of 0.025 to the right, therefore an area 0.975 to the left)

Hence, the 99% confidence interval is

0 .3 0.3
26 − (1.96 ) < u < 2.6 + (1.96 )
36 36

2.5 < u < 2.7

2)99% C.I: α = 0.005 Z 0.005 = 2.575


2
α = 0.01
(Leaving an area of 0.005 to the right and 0.995 to the left)

Hence, the 99% confidence interval is

19
0. 3 0.3
26 − ( 2.575 ) < u < 2.6 + ( 2.575 )
36 36
2.47 < u < 2.73

* we are 95% confidence that the sample mean X = 2.6 differs from the true mean u
by amount less than 0.1 and 99% confident that the different is less than 0.13

One sided confidence bounds on u : σ known

If X is the mean of a random sample of size n from a population with variance σ 2


,the one-sided confidence bound for u are given by

Upper one-sided bound: X + Zα σ


n

σ
Lower one-sided bound: X − Zα
n

Example 2

In a psychological testing experiment, 25 subjects are selected randomly and their


reaction time, in seconds, to a particular experiment is measured. Past experience suggest
that the variance in reaction time to this type of stimuli are 4sec 2 and that reaction time is
approximately normal. The average time for the subjects was 6.2 seconds. Give an upper
95% bound for the mean reaction time.

Solution

The upper 95% bound is given by α = 0.05 (leaving an area 0.05 to the
right, − 0.95 to the left)
σ
x + Zα = 6.2 + (1.645 ) 4 25
n

= 6.858 seconds.

Hence, we are 95% confident that the mean reaction is less than 6.858 seconds.

The number X of successes in n Bernoulli trials is called a binomial random variable.

The probability distribution of this discrete random variable is called the binomial
distribution and its value will be represented by b( x; n, p )

20
A Bernoulli trial can result in a success with probability p and a failure
with probability q =1 − p .Then the probability distribution of the binomial random
variable X , the number of successes in n independent trials,is

n 
b( x; n, p ) =   p x q n −x x = 0,1, n
x

The mean and variance of the binomial distribution are

u = np σ 2 = npq q =1 − p

Example 1
The probability that a certain kind of component will survive a shock test is 3 / 4 .Find
the probability that exactly 2 of the next 4 components tested survives.

Solution
n=2 n=4 p= 3
4
b( x, n, p ) = b( 2,4,3 / 4)

= 4
 
2
( 4) ( 4)
3
2
1
2
= 4!  32
 4
2!2! 
4
 27

 = 128

Using the binomial table

r
Binomial sum : B (r , n, p )
 = ∑ b( x, n, p)
x =0
r −1
1. p ( X < r )
= ∑ b( x, n, p).
x =0

r r −1
2. p ( X = r )
= ∑ b( x, n, p) ∑ b( x, n, p).
x =0 x =0

b
3. p (a ≤ X ≤ b) = ∑ b( x, n, p)
x =a

21
b a −1

= ∑ b( x, n, p) ∑ b( x, n, p).


x =0 x =0

4. p ( X ≥ r )
= 1 p( X < r )
r −1

= 1 ∑ b( x, n, p).
x =0
Example 2

The probability that a patient recovers from a rare blood disease is 0.4. If 15 people are
known to have contracted this disease, what is the probability that
a) at least 10 survive
b) from 3 to 8 survive
c) exactly 5 survive

then, find the mean and variance of the binomial random variable

solution
9
a) p( X ≥10 )
= 1 − P ( X <10 ) = 1 − ∑ b( x,15 ,0.4)
x =0

= 1 − 0.9662
= 0.0338 8
p (3 ≤ X ≤ 8) = ∑ b( x;15 ,0.4)
b)
x =3

8 2

= ∑ b( x;15 ,0.4)
x =0
∑ b( x;15 ,0.4)
x =0

= 0.9050 − 0.0271
= 0.8779 .

(c) p ( X = 5)
= b(5,15 ,0.4)

5 4

= ∑ b( x;15 ,0.4
x =0
∑ b( x;15 ,0.4)
x =0

= 0.4032 − 0.2173
= 0.1859 .
Mean,
= u = (15 ) (0.4) = 6 variance = σ2 = (15 )( 0.4)( 0.6) = 3.6

22
Example 3

It is conjectured that an impurity exists in 30% of all drinking wells in a


certain rural community 10 were randomly selected for testing.
a) Using the binomial dist, what is the probability that exactly three wells
have the impurity?
b) What is the probability that more than three wells are impure?

Solution
3 2
(a) b(3,10 ,0.3)
= p ( X = 3) = ∑ b( x;10 ,0.3) − ∑ b( x;10 ,0.3)
x =0 x =0

= 0.6496 − 0.3828 = 0.2668 .

(b) p( x > 3)
= 1 − p( x ≤ 3)
= 1 − 0.6496 = 0.2668 .

2.3.3. Poisson Distribution

In binomial distribution, the number of successes in n trials is determined. Whereas in


Poisson distribution, the number of successes at a random point of time or space is
determined

For a given time interval, the random variable X may represent

The number of telephone calls per hour


The number of days school is closed during the winter.

For a specified space, the random variable X may represent

The number of bacteria in a given culture


The number of typing errors per page

The probability distribution of the Poisson random variable X representing


the number of outcomes occurring in a given interval or specified region
stated by t is,
p ( x; λ)
e −λ (λ) x x = 0,1
=
x!
Where λ is the average number of outcomes per unit time, distance, area or
volume and
e = 2.71828 ...

Both mean and variance of the Poisson distribution p ( x, λ)

23
u =σ2 = λ

Example 1

During a laboratory experiment, the average number of radioactive particles passing


through a counter in 1 millisecond is 4. What is the probability that 6 particles enter the
counter in a given millisecond?

Solution

x =6 λ =4

e −4 46 6 5
p ( x = 6) = p(6,4) = = ∑ p ( x;4) − ∑ p ( x;4)
6! x =0 x =0

= 0.8893 −0.7851
= 0.1042 .

Example 2

Ten is the average number of oil tankers arriving each day at a certain port city. The
facilities at the port can handle at most 15 tankers per day. What is the probability that on
a given day tankers have to be turned away?

Solution
Let X be the number of tankers arriving each day.

P( X >15 ) =1 − P( X ≤15 )
15
= 1 − ∑ P ( x;10 )
x =0

= 1 − 0.9513
= 0.0487 .

2.4 Some Continuous Probability Distributions

2.4.1 Continuous Uniform Distribution

This distributions is characterized by a density function that is ‘flat’ and thus the

probability is uniform in a closed interval, say [A,B].

24
The density function of continuous uniform random variable x on the interval

[A,B] is

 1
 B − A, A ≤ X ≤ B
f ( x; A, B) = 
 0, e l s e w h e r e

Example

The density function for a uniform random variable on the interval [1,3]

1 1
f ( x;1,3) = = 1≤ x ≤3
B−A 2

0
elsewhere

2.4.2 Normal Distribution

- The most important continuous probability distribution

- Also referred as the Gaussian distribution

- It graph, called the normal curve is the bell shaped curve.

25
The density of the normal random variable X , with mean u and variance σ2 , is
1
1 ( x − u )2
n( x; u ,σ ) = 2σ 2
2πσ e− ,− ∞ < x < ∞
Where
π = 3.14159 e = 2.71828 ...

Fig 2:(a)

26
Fig 3:(b)

Fig 4:(c)

a) Normal curve with u1 < u2 and σ1 = σ 2


b) Normal curve with u1 = u2 and σ1 < σ 2
c) Normal curve with u1 < u2 and σ1 < σ 2

The probability that the random variable X assumes a value between x = x1 and
,
x = x2 i.e.

P( x1 < X < x2 )
is represented by the area of the shaded region.

27
P( x1 < X < x2 )
Fig 5: = area of the shaded region

The transformed normal distribution.

Due to a hopeless task to attempt to set up separate tables for every conceivable values of
u and σ , we use the transformed normal distribution.
X σ 2 =1
Normal random variable, normal random variable, Z with u=0 and

transform

The transformation,
X −u
Z=
σ

Fig 6:The original and transformed normal distributions

28
Example 1
P ( x1 < Xx 2 )
= P ( Z1 < Z < Z 2 )

Given a standard normal distribution, find the area under the curve that lies
(a) to the right of Z = 1.84
(b) between and
Z = −1.97 Z = 0.86 .

Solution
(a) P( Z >1.84 ) =1 − 0/ (1.84 ) =1 − 0.9671 = 0.0329

(b) P (−1.97 < Z < 0.86 ) = 0


/ (0.86 ) − 0
/ (−1.97 )
= 0.8051 −0.0244
= 0.7807

Example 2

Given a standard normal distribution, find the value of right of K such that

(a) P ( Z > K ) = 0.3015

(b) P ( K < Z < −0.18 ) = 0.4197

Solution

29
(a) value leaving the area of to the right, must then leave an area of
K 0.3015
0.6985 to the left

∴K = 0.52

(b) / ( −0.18 ) − 0
0 / ( K ) = 0.4197

/
0 0/ ( K ) = 0.4286 − 0.4197

∴K = −2.37 .

Example 3

Given a random variable X having a normal distribution with u = 50 and σ =10 , find
the probability that X assumes a value between 45 and 62 .

Solution

The value corresponding to x = 45 and x = 62 are


Z 1 2

45 − 50 62 − 50
Z1 = = −0.5 Z2 = = 1.2
10 10

Therefore,
P ( 45 < X < 62 ) = P ( −0.5 < Z <1.2)
=0
/(1.2) −0 /( −0.5)
=0.8849 −0.3085
=0.5764 .

Example 4

30
Given a normal distribution with and σ = 6, find the value of x that has
u = 40

(a) 45% of the area to the left

(b) 14% of the area to the right

Solution

(a) P ( Z < −0.13 ) = 0.45 →Z = −0.13


Hence, x = (6)( −0.13 ) + 40 = 39 .22

(b) P( Z <1.08 ) = 0.86 → Z =1.08


Hence, x = (6)(1.08 ) + 40 = 46 .48

Question 1

A certain type of storage battery lasts on average 3.0 years with a standard deviation of
0.5 year. Assuming that the battery lives are normally distributed, find the probability
that a given battery will last less than years.
2. 3

Ans: P ( X < 2.3) =


0.0808
X −U
Z=
σ

31
Question 2

An electrical firm manufactures light bulbs that have a life before burn-out, that is
normally distributed with mean equal to 800 hours and a standard deviation of 40 hours.
Find the probability that a bulb burns between 778 and 834 hours.

Ans: P(778 < X < 834 )


0.5111

Question 3

The average grade for an exam is 74 , and the standard deviation is 7. if 12% of the
class is given A’s and the grades are curves to follow a normal distribution, what is the
lowest possible A and the highest possible B?

Ans: lowest A is 83 , highest B, is


82

Normal Approximation to the Binomial

Let X be a binomial random variable with parameters n and p. Then X has


approximately a normal distribution with u
u = np ,
σ 2 = npq = np (1 − p)
and

x
P( X ≤ x) = ∑ b ( k ; n, p )
k =0
Area under normal curve to the left of
≈ x + 0 .5

=  x + 0.5 − np
P( Z ≤ 

 npq



Example 1 b( x; n, p )) = b( x;15 ,0.4)

Using binomial distribution for P( X = 4)


P ( X = 4) = b(4;15 ,0.4)

32
4 3

= ∑ b( x;15 ,0.4) − ∑ b( x;15 ,0.4)


x =0 x =0

= 0.1268 .

Using normal distribution for P ( X = 4)

u = np =
15 (0.4) = 6
Approximate area between x = 3.5 and
1 x2 = 4.5
σ 2
= npq = 15 (0.4) (0.6)3.6 =
3.5 − 6
Converting to Z values, Z1 = = −1.32
1.897
4.5 − 6
Z2 = = −0.79
1.897

Hence, P( X = 4) = b(4,15 ,0.4)


≈ P (−1.32 < Z < −0.79 ) = P(Z < −0.79 ) − P(Z < −1.32 )
= 0.2148 −0.0934
= 0.1214

Example 2 P (7 ≤ X ≤ 9)

Using binomial dist:


9 6
P (7 ≤ X ≤ 9) = ∑ b( x;15 ,0.4)
x =0
∑ b( x;15 ,0.4)
x =0

= 0.9662 − 0.6098
= 0.3564 .

Using normal dist:


Approximate area between and x = 9.5
u = 6 σ = 3.6 x1 = 6.5 2

6.5 − 6
Converting Z values, Z1 = = 0.26
1.897

9.5 − 6
Z2 = = 1.85
1.897
Hence, P (7 ≤ X ≤ 9)
≈ P (0.26 < Z <1.85 )

33
= P ( Z <1.85 ) − P ( Z < 0.26 )
= 0.9678 − 0.6026
= 0.3652 .

2.4.3 The exponential distribution

Exponential distribution is the special gamma distribution for which x =1.

The continuous random variable x has exponential distribution with


parameters B, if its density function is given by
1 −xB
f ( x; B ) = e ,x > 0
B

0, elsewhere
Where B > 0

The mean and variance of exponential distribution

and
u =B σ 2 = B2

P ( Z < 0)

=φ(0)
=0.5

34
P ( Z <3.30 )
=φ(3.3)
=0.9995

P ( Z <0.7 )
=φ(0.7)
=0.7580

P ( Z <−0.1)
=1 −φ(0.1)
=1 −0.5398
=0.4602

P ( Z >−1.96 )
= P ( Z <1.96 )
=φ(1.96 )
=0.9750

P ( Z >−0.66 )
=1 −φ(0.66 ))
=1 −0.7454
=0.2546

35
P ( 0 < Z < 2)
=φ( 2) −φ(0)
=0.97725 −0.5
=0.4772

P (1.5 < Z < 2.7)


=φ( 2.7) −φ(1.5)
=0.9965 −0.9332
=0.0633

P ( −2.67 < Z < 2)


=φ( 2) −φ( −2.67 )
=φ( 2) −[1 −φ( 2.67 ]
=0.9772 −[1 −0.9962 ]
=0.9734

P ( −2.7 < 2 <1.5)

3.1 Population, samples and sampling distributions

Population is the set of numbers, measurements or observation which are of interest.

36
- size of the population N is the number of object or observation in the population
- population is said to be finite or infinite depending on the size of N being finite
or infinite

Sample is a finite subset of the population

-size of sample is represented by n

Parameter is a statistical measure obtained from the population

-e.g population mean, population variance

Statistic is a statistical measure obtained from the sample

--e.g sample mean, sample variance

Probability distribution of a statistic is called sampling distribution

Notation

Population Sample
Mean u x
Standard deviation σ s
proportion p P

Population f ( x) = a population whose probability distribution is e.g is f ( x ) e.g. if


f ( x ) is binomial, poisson or normal, then the corresponding population is known as
binomial population, Poisson population or normal population.

In sampling distribution, its statistic is a random variable.

Sample mean and sample variance are two important statistics with are measure of a
random sample X 1 , X 2 , X n of size n
(measure of central tendency)
n

Sample mean ∑X i
(measure of central tendency)
=X = i =1
n

∑( X )
n

i −X2
Sample variance (measure of variability
=s = 2 i =1
n −1
of data about the mean)

37
n∑ X i2 − ( ∑ X i )
2

=
n( n −1)

Sample standard deviation, S - a positive square root of the sample variance.

3.2 sampling distribution of mean ( σ known)

Suppose that a random sample of n observation is taken from a normal population with
mean, u and variance σ 2

Each observation X i , i =1,2, n of the random sample will have the same normal
distribution. The normal approximation of the sample mean, X can be calculated using:

The standard normal distribution n( z;0,1)


uX = u 2 σ2
σ =
X n

X −u
Z=
σ/ n

Where n –sample size, u and σ are mean and variance from the population and X is
mean of random sample

Example 1

An electrical firm manufactures light bulbs that have a length of life that is approximately
normally distributed, with mean equal to 800 hours and a standard deviation of 40 hours.
Find the probability that a random sample of 16 bulbs will have an average life of less
than 775 hours.

Solution

Sampling distribution of will be approximately normal, with u x = 800 and


40
σX = = 10
16

Hence, corresponding to X =775 ,


775 − 800
Z= = −2.5
10

Therefore, P( X < 775 ) = P ( Z < −2.5) = 0.0062

38
What is the t – Distribution used for?
- to deal with inference about the population mean. ( S as estimator of σ)
-will be explained in the next chapter.

Example 3

A chemical engineer claims that the population mean yield of a certain batch process is
500 grams per mililiter of raw material. To check this claim, he samples 25 batches each
month. If the computed t –value falls between −t0.05 and t0.05 , he is satisfied with
his claim.
What conslusion should be draw from a sample that has mean x = 518 grams per mililiter
and a sample standard deviation, S = 40 grams? Assume the distribution of yields to be
approximately normal.

Solution: ANS: the process produces a better product than he thought.

3.3 Sampling distribution of mean ( σ unknown) : t -distribution

Earlier in problems of inference on a population mean, it was assumed that the population
standard deviation, σ is known.

However, in many experiments, the knowledge of σ is no more reasonable than


knowledge of the population mean u

Often, in fact an estimate of σ must be suplied by the same sample information that
produced the sample average x

As a result, a natural statistic to consider to deal with inference on u is


X −u
T = With r = n −1 degrees of freedom
S/ n

• if the sample size is large enough, say n ≥ 30 ,the distribution of T does not
differ considerably from the standard normal.
• However, for n < 30 , it is useful to deal with the exact distribution of T
- in developing the sampling distribution of T ,we shall assume that our
random sample was selected from a normal population.

To sum-up-use the table below

n ≥ 30 n < 30 (*)replace σ by standard


σ known Z Z deviation sample, S.
σ unknown Z (*) T 1 n
S2 = ∑ ( xi − x ) 2
n − 1 i −1
X −u
Z=
σ/ n

39
Further explaination on sampling distribution

1. Draw all possible samples of size n from a given finite population of size N
Total number of possible samples of size n from the population of size N is given

N1
NCn = 0
=K
n1( N − n ) 1
0 0

2. Compute a statistic S (such as mean,s.d, median, mode etc) for each of these samples.
S = S ( x1 , x2 , x3  xn )

Sample number 1 2 3 K
Statistic S S1 S2 S3  S K

- Thus, sampling distribution describe how the statistic S will vary from one sample to
the other of the same size.
- The difference in the value of the statistic is known as sampling fluctuations.
- If the statistic S is mean, the sampling distribution is known as sampling distribution of
means.

S Sampling distribution
Mean Sampling dist of mean
Variance Sampling dist of variance
Proportion Sampling dist of proportion
Median Sampling dist of median


40
4.1 Statistical inference

Statistical inference is a method by which one make inferences or generalisations about a


population by using the knowledge of sampling distribution.

Statistical Inference

Estimation Test of Hypothesis

Bayesian method

Interval estimate

Point estimate

Estimaton – judgement or opinion of the population parameters.

Test of hypothesis-Decision in acceptance or rejection of statements about the population


parameters.

4.2 Estimation and Test of Hypothesis

Case study-Estimation

A candidate for president post may wish to estimate the true proportion of voters favoring
him by obtaining te opinions from a random sample of 100 eligible voters. The fraction
of voters in the sample favoring him could be used as an estimate of the true proportion
in the population of voters.
To sum up,

41
If X is the mean of a random variable sample of size n from a population with known
variance σ2 , a
α = 0.05
α = 0.01
100 (1 − α ) % Confidence interval

For u is given by

X − Zα σ σ
< u < X + Zα
2 n 2 n

Where Zα 2 is the Z -value leaving an area of α to the right


2

What does 100 (1 − α ) % mean?

-when α = 0.05 , we have a 95% confidence interval. (we can be 95%confident that the
given interval contain the u )

-Then when α = 0.01 , we obtain a wider 99% confidence interval. The wider
confidence interval, the more confident we can be that the given interval contain the u.

Error

X − Zα α X u X + Zα 2 α n
2 n

The size of error is the absolute value of the difference between u and X
We can be 100 (1 − α ) % confident that this error will not exceed Zα 2 σ n

4.3. Estimation
4.3.1 Point estimation
~ as tought in chapter 3, (3.2 3.5) ~

4.3.2 Confident interval of u with variance known


Let us consider the interval estimate of u . If our sample is selected from normal
population or failing this, if n is sufficiently large, we can establish the

100 (1 − α )% Confidence interval

42
For u by considering the sampling distribution of X

By refering figure 1 below, writing Zα 2 for Z - value,we can find an area of α 2

1 −α

1 :degree of confidence

We then interpret figure 1 as

X −u
P − Zα < Z < Zα  = 1 − α , Z =
 2 2 α n

 X −u 
P − Zα < < Zα  = 1 − α
 2 α n 2

 α σ 
P X − Zα < u < X + Zα  = 1− α
 2 n 2 n

4.3.3 Confidence interval of u with variance unknown.


Small sample confidence interval for u (n < 30 , assuming sampling from normal
population) with variance unknown, statistic T is then used

This procedure is the same as 4.2.2 section except that σ is replaced by S and the
standard normal distribution ( Z distribution) is replaced by the t –distribution.

(* For large sample ( n ≥ 30 ) with unknown variance, use z dist. The unknown σ is
substituted by s )

Refering figure 2,

43
If X and s are mean and standard deviation at a random sample from a normal
population with unknown variance σ 2 ,a 100 (1 − α ) % confidence interval for u is

S S
X − tα < u < X + tα
2 n 2 n

Where tα 2 is the t –value with degree of v = n −1 degree of freedom

Example

The contents of 7 similar containers of sulfuric acid are 9.8,10.2, 10.4,9.8, 10.0, 10.2
and 9.6 liters. Find a 95% confidence interval for the mean of all such containers,
assuming an approximate normal distribution.

Solution

The sample mean and standard deviation for the given data are

X =10 .0 and S = 0.282

From table, t0.025 = 2.447 for v = 6 V = n −1


Hence, the 95% confidence interval for u is
 0.283   0.283 
10 .0 − ( 2.447 )  < u < 10 .0 + ( 2.447 ) 
 7   7 
9.74 < u < 10 .26
Example- Point estimate

44
The value x of the statistic X in the particular sample and sampling distribution is a
point estimate of a population parameter u

Example –Interval estimate

An interval 0 L < u < 0U is an interval estimate of a population parameter u . 0 L


(lower limit) and 0U (upper limit) depend on the value x of the statistic X in
particular sample and sampling distribution.

Case study- Test of Hypothesis

One is interested in finding out whether brand A floor wax is more scuff resistant than
brand B floor wax. He might hypothesis that brand A is better than brand B and after
proper testing (using particular sample and sampling distribution, he then accept or reject
this hypothesis.

Here, we do not attempt to estimate a parameter but instead we try to arrive at a correct
decision about a prestated hypothesis. As in estimation, sampling theory is also applied
here.

4.4 Test of Hypothesis

Null Hypothesis

- Null hypothesis is tested for possible rejection by assuming (before testing) that it is
true.
- Stated as H 0 and the statement expressed as equality (=)

H 0 : u = u0
i.e we assumes that the population mean equals to u0

Alternative hypothesis

-Any hypothesis other than the null hypothesis.


-stated as H1 and statement expressed as inequalities such as <,> or =
/
H1 : u > u0 or
H1 : u < u0 or

H1 : u =/ u0

Test of hypothesis

45
Test of hypothesis is a procedure to decide whether to accept or reject the null hypothesis
based on statistic in sampling distribution.

Decision
Accept H 0 Reject H 0
H 0 is true Correct decision Type! error
H 0 is false Type II error Correct decision

Critical region

Consider any statistics S ∗ (statistics Z or T for example), the area under the
probability curve can be divided

Into: 1) critical region

2) non – critical region

The critical region is the region of rejection of NULL hypothesis.

Example
Non critical region critical region

Two tailed test


Η0 : µ = µ0 x − µ0
Ζ=
Η1 : µ ≠ µ0 σ/ n
t=

Left one tailed test

Η0 : µ = µ0 Ζ
Η1 : µ < µ0 t

Right one tailed test

46
Η0 : µ = µ0 Ζ
Η1 : µ > µ0 t

4.3.1 Test the hypothesis with known variance σ 2


Left u and σ 2 be the mean and variance of a population from which a random sample of
size n is drawn.
Then, left x be the mean of the sample. For large sample (n ≥ 30 ), the sampling
distribution of x is approximately normally distributed with
Mean, u = u and variance, 2 σ 2 is
x σ =
x n

The test statistics is


x − u0
Ζ=
σ/ n
For large sample, n ≥30 , even if σ is unknown, σ can be replaced by sample
Χ
variances, S .

a. To test whether the population mean u equals to specified constant u0 or not, i.e.
when Η1 : u ≠ u0

1. Η0 : u = u0
2. Η1 : u ≠ u0
3. Choose α
4. since Η1 : u ≠ u0 (two tailed test)
Calculate − Ζα 2 and − Ζα 2
5. Compute test statistics
x − u0
Ζ= =
σ/ n
6. Decision: reject Η0 if Ζ < − Ζα or Ζ > Ζα
2 2

Otherwise accept Η0 if − Ζ α < Ζ < Ζ α


2 2
b. when Η1 : u < u0
1. Η0 : u = u0
2. Η1 : u < u0
3. Choose α

47
4. since Η1 : u < u0 (left one tailed test)
Calculate − Ζα
5. Compute test statistics
x − u0
Ζ=
σ/ n
6. Decision: reject Η if Ζ < −Ζ
0 x

Otherwise accept Η if Ζ > −Ζ


0 x

c. when Η1 : u > µ0
1. Η0 : u = u0
2. Η1 : u > u0
3. Choose α
4. since Η1 : u > u0 (right one tailed test)
5. Compute test statistics

x − u0
Ζ=
σ/ n
6. Decision: reject Η if Ζ > −Ζ
0 α

Otherwise accept Η if Ζ < −Ζ


0 α

Example 1

A manufacture of sports equipment has develop a new synthetic fishing line that he
claims has a mean breaking strength of 8 kilograms with a standard deviation of 0.5
kilograms. Test the hypothesis that u = 8 kilograms against the alternative that u ≠ 8
kilograms if a random sample of 50 lines is tested and found to have a mean breaking
strength of 7.8 kilograms. Use a 0.01 level of significance.

Solution

1. Η 0 : u = 8 kg

2. Η1 : u ≠ 8 kg
3. α = 0.01

4. Critical region: Zα 2 = −2.575,2.575

5. Compute test statistics

48
x = 7.8 kg , n = 50 , u =8 kg

σ =0.5 kg .
x − u0 7.8 − 8
Z= = = −2.83
σ n 0.5 50

6. Decision:

Ζ < − Ζα : − 2.83 < 2.575


2

α = 0.01
Reject and conclude that the average breaking strength is not equal to 8 but in

fact less than 8 kg.

Example 2

A random sample 0f 100 recorded deaths in the United States during the past year
showed an average life span of 71.8 years. Assuming the population standard deviation of
8.9 years, does this seem to indicate that the mean life span today is greater than 70 years
use a 0.05 level of significance.

Solution
1. Η0 : u = 70 years
2. Η1 : u > 70 years
3. α = 0.05
4. Critical region: Ζα =1.645

5. Compute test statistics,


x = 71 .8, u = 70 , σ = 8.9, n =100

49
x − µ0 = 71 .8 − 70 = 2.02 .
Ζ=
σ / n 8.9 / 100

6. Decision:

Ζ > −Ζα : 2.02 >1.645


Reject Η0 and conclude that the mean life span today is greater than 70 years

Example 1

An electrical company claimed that a vacuum cleaner expends an average of 46


kilometers hours per year. If a random sample of 12 homes included in a planned study
indicates that vacuum cleaners expand an average of 42 kilowatt hours per years with a
standard deviation of 11.9 kilowatt hours, does this suggest at the 0.05 level of
significance that vacuum cleaners expand, on the average, less than 46 kilowatt hours
annually? Assume the population of kilowatt hours to be normal.

Solution
1. Η0 : u = 46 Kilowatt hours
2. Η1 : u < 46 Kilowatt hours
3. α = 0.05
4. Critical region: tα = −1.796 with 11 degrees of freedom
5. Compute statistics T

x = 42, S =11 .9, n = 12

50
Hence,

x − u0 42 − 46
t= = = −1.16
S / n 11 .9 / 12

6. Decision:

t > tα : −1.16 > −1.796


Do not reject Η0 and conclude that the average number of kilowatt hours is not
significantly less than 46.

4.3.2. Test the hypothesis (small sample and σ unknown − t distribution

For σ unknown and for small sample size, the decision criterion is based on the t
distribution with V = n −1 degrees of freedom.

The test statistics is

x − u0
t=
S/ n

The test is similar as in section 4.3.1 except t values are used in place of Ζ and σ is
replaced by S

5.1 Curve fitting

5.1.1 Introduction

51
Curve fitting is a method of determining the mathematical equation between the variables
X and Y by finding the best fit curve, C : Y = f ( x ) that approximates a given set of
data.

Relationship

- The relationship between two variables, X and Y (or more) may exist. For
example : blood pressure and age, nutrition and IQ etc.

- The relationship of the variables can be ilustrated in a scatter diagram.

Scatter diagram

- Scatter diagram is the plotted points of the set of given N paired observation of X
and Y , i.e ( x1 , y1 ), ( x2 , y2 )  ... ( xn , yn ) in XY plane
plane

Figure 1: scatter diagram

Approximating curves, C

- A smooth curve that approximates the given set of N data points plotted in the scatter
diagram.

Best fitting curve by least square

- The use of least square method to find the approximating curve, C

52
-Let ( x1 , y1 ), ( x2 , y2 ), ( xn , yn ) be a given set of N data points

d i =Yi −Yˆi Denotes the difference between Yi and the corresponding value Yi
estimate
Yˆi = f ( X i )

N
- The curve that having the minimum ∑d
i =1
i
2
is the best fitting curve and

∑d
i =1
i
2
is called the residuals or error sum of squares.

-Such curve is known as a least square (L.S) curve.

Some standard approximating curve

1. Y = a0 + a1 X (straight line)
2. Y = a0 + a1 X + a2 X
2
(quadratic curve)
3. Y = A B* (exponential curve) etc.

5.1.2 Curve fitting by least square

For a given set of N data points ( X 1 , Y1 ), ( X 2 , Y2 ), ( X N , YN ) , assume that the


straight line

Y = a0 + a1 X = f ( X )

Fit to the data in the least square sense

The two unknown parameters a0 and a1 are determined from

∑Y i = Na 0 + a1 ∑X i

∑X Y i i = a0 ∑X i + a1 ∑X i2

Example 1

Find a least squares straight line for the following data:

X : 1 2 3 4 5 6

53
Y:6 4 3 5 4 2

And estimate (predict) Y at X = 4 and X at Y = 4

Solution

∑X =21 ∑Y =24 ∑X 2
= 91 ∑Y 2
=106 ∑XY =75

n =6

Assume that the least squares straight line of Y on X is Y = a0 + a1 X

∑Y = Na 0 + a1 ∑X a1 =

∑XY = a0 ∑X + a1 ∑X 2 a0 =

Substituting the values

24 = 6a0 + 21a1

75 = 21 a0 + 91 a1

Solving a0 = 5.7999 , a1 = −0.5143

Thus the least square straight line Y on X is

Y = 5.7999 − 0.514 X

Y estimate = Y (at ( X = 4 ) ) = 5.7999 − 0.514 ( 4)


= 3.743

5.2 Regression analysis

- The study of the relationship (if it exists) between two (or variables).
- Mainly used for

1) Prediction- to estimate one dependent variable (response) in terms of the other


independent variable (s) (regressor (s))
2) Optimization- to determine the values of independent variables (s) for which the
dependent variable attains maximum or minimum.

54
- Linear regression – the relationship between the variables is linear. Represented by a
straight line.
If Y is dependent on X then Y = a0 + a1 X is known as regression line of Y on X
Similarly, if X depends on Y , then X = b0 + b1Y is known as regression line of X
on Y

Simple linear regression model

Simple linear regression model consist of

Y = α + βx + ε
α -unknown intercept
β-unknown slope
ε - random error (assume to be normally distributed with mean E (ε ) = 0 and variance
σ 2 (residual variance)

To estimate the α and β, a regression line Yˆ = a0 + a1 X is fitted according to principle


of least square.

Similiarly, least square straight line of X on Y is assumed to be

X = b0 + b1Y

∑X = Nb 0 + b1 ∑Y

∑XY = b0 ∑Y + b1 ∑Y 2

Or 21 = 6b0 + 24 b1

75 = 24 b0 +106 b1

Solving b0 = 7.1 , b1 = −0.9 so


X = 7.1 − 0.9Y

X estimate = X (at Y = 4 ) = 7.1 −0.9( 4 )


= 3.5

5.3 Correlation Analysis

- In the regression analysis, the relation in the form of mathematical equation is obtained.
- While in correlation analysis, we wish to find whether the relationship exist and the
strenght of the relationship is measured.

Regression equation-used for prediction

55
Correlation coefficient-measure the closeness

Type of correlation

By plotting a given set of n pairs of random variables ( X i , Yi ) , for i =1,2,3, as a


scatter diagram, the correlation is said to be

For the linear correlation coefficient, r


i) r lies in the interval [−1,1] , i.e − 1 ≤ r ≤ 1
ii) r is independent of origin
iii) r is independent of unit

 n  n 
n
 ∑ x  ∑ y 
∑ xy −  i =1  i =1 
n
r= i =1

  n  n    n  n  
 n  ∑ x  ∑ x   n  ∑ y  ∑ y  
  i =1  i =1   
∑x −
2
  ∑ y −  i =1  i =1  
2

n n
 i =1  i =1 
  

56
Example 1

Find
a) t0.025 when v = 14
b) −t0.01 when v = 10
c) t0.995 when v = 7
d) t − value which v = 14 that leaves an area of 0.025 to the left

solution:

a) t0.025 = 2.145
b) −t0.01 = 2.764
c) t0.995 = t1−0.005 = −t0.005 = −3.499
d) t0.975 = −t0.025 = −2.145

Example 2

Find k for a random sample of size 24 from a normal distribution such that

a) P( − 2.069 < t < k ) = 0.965


b) P( k < t < 2.807 ) = 0.095
c) P( − k < t < k ) = 0.90

Solution

a) tα = 2.069 with v = 24 −1 = 23 ,α = 0.025

the area to the right of


k = 1 − 0.965 − 0.025 = 0.01
∴k = t0.01 = 2.50 with v = 23

57
b) p( k < t < 2.807 ) = 0.095
tα = 2.807 with v = 23 , α = 0.005
Area to the left of k is 1 − 0.005 − 0.095 = 0.9
Then, area to the right of K is 0.1
∴K = t0.1 =1.319 with v = 23

c) P (−K < t < K ) = 0.90

Since 1 − 2α = area given = 0.9


0 .1
α= = 0.05
∴ 2

∴K =t 0.0.5 =1.714 with v = 23

Example 2

A process for making certain ball bearings is under control if the diameters of the
bearings have a mean of 0.5 cm. If the random sample of 10 of these bearings has a mean
diameter of 0.5060 cm and s.d of 0.004 cm, is the process under control? (use α = 0.005
)

Solution

X = 0.5060 (sample mean), u = 0.5 (population mean), v = 9 n = 10


s = 0.004 (sample s.d)

Then,

x −u
t= = 4.7434 since 4.7434 > tα = t0.005 = 3.250
s n
the process is not under control.

z
t
t dist z dist
Mean 0 0
variance 1 More than 1(depend on v = n −1)
as n → ∞ , variance approaches 1

58
Critical values of t dist

Critical value, tα : the area under the curve to the right of tα equal to α

Symmetry property

t1−K = −tα

e.g t0.95 = −t0.05


t0.99 = −t0.01

Confidence interval

X − Zα σ σ
< u < X + Zα
2 n 2 n

X − tα s σ
< u < X + tα
2 n 2 n

59
Test of hypothesis –to conclude

1.
2.
3.
4.calculate: if two sided
− Zα and Zα (or − tα and tα )
2 2 2 2
If one sided left if one sided-right
− Zα (or −tα ) Zα (or  tα )

5. Compute test statistic

x − u0 x − u0
Z= or t=
σ n σ n

6. Conclusion

Reject H 0 /Do not reject H 0


Use alternative hypothesis statement to conclude

60
61

Das könnte Ihnen auch gefallen