ENGINEERING MATHEMATICS II
010.141
STATISTICAL
NUMERICAL SOLUTIONS
MODULES 7
STATISTICAL NUMERICAL SOLUTIONS
Ø 24.1 Data
Ø 24.2 Experiments, outcomes, events Ø 24.3 Probability
Ø 24.5 Random variables, and Probability distribution Ø 24.6 Mean and variance
Ø 24.7 Binomial, Poisson, and Hypergeometric Ø 24.8 Normal
Ø 24.9 Distribution of several random variables
DATA REPRESENTATION
Ø Average; spread Ø Histograms
Ø Center; median; quartile Ø Box plot
Ø Outlier
Ø Mean; standard deviation; variance.
ILLUSTRATIONS
Ø Consider these 14 umbers:
78, 81, 83, 84, 86, 87, 87, 89, 89, 89, 89, 90, 91, 99
Ø These are arranged in order ; half-way is between 87 and 89; the median is half-way, so the median would be either 87 or 89.
Ø The numbers could be grouped as follows:
§ A: 77 +/- 2.5 A: one number
§ B: 82 +/- 2.5 B: 3
§ C: 87 +/- 2.5 C: 7
§ D: 92 +/- 2.5 D: 2
§ B: 97 +/- 2.5 E: 1
EXAMPLE OF HISTOGRAM
Number by interval
A B C D E
6 7
4 5
2 3
0 1
SOME DEFINITIONS
Ø Absolute frequency: number of times a value shows up
Ø Relative frequency: number of times a value shows up, divided by total of numbers
Ø Cumulative absolute frequency: running total Ø Range: from min value to max value
Example: 78 to 91 is a range of 21
Ø Median: data value that falls in the middle, when lined up in order;
in current example, between 87 and 89; use 88 as median
Ø Outlier: value that seems to be different from the other values in the set.
MORE DEFINITIONS
Ø Mean: average of the data; sum of the values, divided by the number of values.
Indicated by (x - bar) Sometimes called the arithmetic mean
In the running example, there are 14 numbers; the sum of the values is 611, so the mean is 611/14, or 87.3.
Note that the mean is slightly below the median, 88
VARIANCE AND STANDARD DEVIATION
Ø The variance indicates spread in the data
Ø The definition of variance of a data set, labeled s2 is:
Ø In the current example, s2 = 25.14
Ø The standard deviation, s, is: s = 5.014
( )
s = 1
n 1 x x
2
j j = 1
-
å
- 2n
SECTION 24.2
EXPERIMENTS, OUTCOMES, EVENTS
Ø Trial: single performance of an experiment Result is an outcome or sample point n trials yields a sample of size n
Sample space is the set of all possible outcomes
Ø Set union: The union of two sets is a set whose elements are in either set, or both sets
Ø Set intersection: The intersection of two sets is a set whose elements are in both sets (intersection might be empty).
Ø Set complement: another set whose elements are in the sample space, but not the first set.
ROLL OF THE DICE
Ø Two dice are rolled; there are six numbers, 1 to 6, on each; the sum of the numbers could be as low as 2 or as high as 12. The dice are assumed to be fair. The next chart depicts the 36 outcomes. Note that the number 7 shows up as the sum six times, and the number 11 shows up twice. If either 7 or 11 is a winner on the roll, then the likelihood of winning on this first roll is 8/36, or 22%.
Suppose that rolling 2, 3, or 12 is considered a loser. There are four such cases in the table. The likelihood of losing on the first roll is thus 4/36, or 11%.
Note that the likelihood of winning is twice losing, on the first roll at least.
OUTCOMES OF ROLLS OF DICE
Ø Roll of two dice and sample space of outcomes of sum of two dice.
1 2 3 4 5 6
1 2 3 4 5 6 7
2 3 4 5 6 7 8
3 4 5 6 7 8 9
4 5 6 7 8 9 10
5 6 7 8 9 10 11
6 7 8 9 10 11 12
SECTION 24.3 PROBABILITY
Ø Definition 1: probability
§ Pr(A) = (number of points in A) / (number of points in S)
§ Pr(S) = 1
§ 0 ≤ Pr ≤ 1
Ø Relative frequency of A:
§ frel(A) = f(A)/n
§ relative frequency lies between 0 and 1
§ relative frequency of S is 1
§ if two sets A and B are mutually exclusive (intersection is empty) then relative frequency of the union of A and B is the sum of their relative frequencies
Pr(A È B) = Pr(A) + Pr(B)
PROBABILITY
Ø Definition 2. Given sample space S; event A in S (A is a subset of S), there are some axioms of probability:
§ 0 ≤ Pr ≤ 1
§ Pr(S) = 1
§ Pr(A È B) = Pr(A) + Pr(B),
for the case where A and B are non-overlapping (that is, their intersection is null)
ADDITION RULE
Ø Given events A and B in a sample space
Pr(A È B) = Pr(A) + Pr(B) - Pr(A Ç B) Note that if A and B are exclusive, the last term is 0 The probability of the null set is zero
Conditional probability of B, given A:
Pr(B|A) = Pr(B Ç A) / Pr(A) where Pr(A) not = 0
( ) ( )
( ) ( )
Pr A|B = Pr A B
Pr BI Pr B 0
; ¹
MULTIPLICATION RULE
Suppose in a sample space S there are two events A and B and their probabilities are greater than zero.
Then:
Pr(A Ç B) = Pr(A) · Pr(B|A) = Pr(B) · Pr(A|B) Conditional Probability is labeled A|B
Ø Independence: If A and B are independent, then:
Pr(A Ç B) = Pr(A) · Pr(B)
SECTION 24.5
Ø Define:
F(x) = Pr(X ≤ x) F(x) is called a distribution function
Pr(a < X ≤ b) = F(b) - F(a) Ø A probability function could be defined:
Density function
In this case it is called discrete; the distribution function is thus:
RANDOM VARIABLES
PROBABILITY DISTRIBUTIONS
( )
f x pj x
= = x otherwise
j
0 ìí î
( ) ( )
F x =
å
f x = jå
pjA random variable X and its distribution function F(x) is continuous provided it can be represented by:
the function f is referred as the density or density function;
As before,
It is necessary that the integral over all possible values be unity.
CONTINUOUS
( ) ( )
F x = f d
x
n n
-¥
ò
( ) ( )
Pr a < x b = f d
a b
£
ò
n n( ) ( )
f x = F x¢
SECTION 24.6
MEAN AND VARIANCE OF A DISTRIBUTION
Consider a uniform distribution (see next page), where f(x) = 1/(b - a), on a < x < b; otherwise = 0.
( )
( )
( ) ( )
( ) ( )
( )
m
m
s m
s m
s s
= x
= x
= x
= x
> 0 Note: could be zero
i i
2
i i
2
f x
f x dx
f x
f x dx
j
j
å ò
å ò
-¥
¥
-¥
¥
-
-
2
2 2
2
UNIFORM DISTRIBUTION
Height = 1/(b - a); Area = 1
The red curve illustrates a more informed density function. The uniform
40 90
70 ?
A B
MEAN AND VARIANCE OF UNIFORM DISTRIBUTION
( ) ( )
( )
( )
( )
b 2 2 2
a
b 2 3
2 a
3 3
3
x x b a a + b
= = = =
b a 2 b a 2 b a 2
a + b 1 1 a + b
= x = x
2 b a 3 b a 2
1 a + b a + b
= b a
3 b a 2 2
1 b a
= 3 b a 2
b
a
b
a
dx
dx m
s
-
- - -
æ ö æ ö
- × × -
ç ÷ ç ÷
- -
è ø è ø
ì ü
ïæ ö æ ö ï
×íç - ÷ - ç - ÷ ý
- ïîè ø è ø ïþ
æ - ö
× ç ÷ -
- è ø
ò ò
( )
( )
( ) ( )
( )
3
3 2 2 3 3 2 2 3
3
2 2
a b 2
= 1 3ab + 3a b a a + 3a b 3ab + b 24 b a
1 2 b a
24 b a b a = 12
b
s
ì - ü
ï æ ö ï
í ç ÷ ý
è ø
ï ï
î þ
× - - - -
-
= × -
- -
SECTION 24.7
BINOMIAL, POISSON,
HYPERGEOMETRIC DISTRIBUTIONS
Ø Binomial: how often does event A occur in n independent trials?
Let p = probability of success per trial and q failure; p + q = 1 (also, q is the same as "not p")
Let X = number of times that A occurs in n trials for X = x, the probability is:
( )
n
x p q = n!
n x p q
x n x x n x
æ èç ö
ø÷
-
- -
x! !
RATIONALE FOR BINOMIAL
For x successes in n trials (n - x) failures
one arrangement is p · p · p · · · p · q · q · · · q with probability px qn-x
for all possible permutations use Theorem 1(b), page 1065:
Probability of x successes in n trials
( )
n!
n x p q = n
x p q
x n x x n x
x! - !
æ èç ö
ø÷
- -
MEAN AND VARIANCE OF BINOMIAL
Ø Mean
μ = np Ø Variance
σ2 = npq If
( )
2
p = q = 1 flip a coin 2
= n 2 = n
4
m
s
DISCRETE POISSON DISTRIBUTION
( )
f x = x e x = 0, 1, 2, = 0! = 1
x 2
m m s m
! - L,
Example 2, page 1081
p = 0.01 n = 100 μ = np = 1 0 ≤ x ≤ 100
Pr .
Pr .
Pr . ob
ob
ob
x = 0 = 1
0! e = e = 36%
x = 1 = 1
1! e = e = 36%
x = 2 = 1
2! e = 1
2 e = 18%
0
1
2
- -
- -
- -
1 1
1 1
1 1
Ø Suppose that the probability that an item produced by a machine will be defective is 0.1. What is the probability that a sample often items, selected at random from the output of the machine, will have no more than one defective?
Answer from binomial:
Answer from Poisson:
EXAMPLE - BINOMIAL VS. POISSON
( ) ( ) ( ) ( )
10
0 01 0 9 10
1 01 0 9
0 10 1 9
æ èç ö
ø÷ æ
èç ö ø÷
. . + . . = 0.7361
( )( )
m = np = 10 0.1 = 1 e + e = 0.7358-1 -1
SECTION 24.8
Ø Definition of normal distribution
Ø Integral form
NORMAL DISTRIBUTION
( )
f x = 1
2 exp x
s p
m - æ s-
èç ö
ø÷ é
ë ê ê
ù û ú ú 1
2
2
( )
F x = 1 d
2 exp
x
s p
n m
s n
- æ - èç ö
ø÷ é
ë ê ê
ù û ú
-¥
ò
21 ú2
PROPERETIES OF THE NORMAL DISTRIBUTION
Ø Mean μ Ø Variance σ2
For standard normal
( ) ( )
f z
z d
z
= 1 2 e
= 1 2
e
x
z
p
f
p
m
m
m -
-
-¥
ò
-
2
2
2
2
EXAMPLE
( )
m s s
m s
m s
= 50 = 9 = 3 Find c such that Pr x < c = 5%
x = c 50
8, page A90 for 5%
x = 1.645
c 50
= 1.645 c = 45.065
2
- -
- -
- -
3
3
See Table 95%
45 47 56
44 x = 50 53
: = 50
EXAMPLE
( )
= 0.01 = 0.001
Probability that R is between 0.009 and 0.011 at lower end
x
= 0.009 0.01
0.001 = 0.001
0.001 = 1.000 x
= 0.011 0.01
0.001 = 0.001
0.001 = 1.000 89
1 = 0.8413 0.1587 = 0.6826 = 68%
For 1000 wires 683
088, figure 489(a)
m s
m s
m s
f
- - -
-
- -
-
® See page A
See page 1
GRAPH OF THE NORMAL DENSITY FUNCTION
( )
( ) ( )
f
p
f
p x = 1
2 e Standard Normal
x = y = 1
2 e dy
x -
-¥
-
-¥
ò ò
1 2
1 2
2
2
x
x y
dy F
GRAPH OF THE NORMAL DENSITY FUNCTION
N(x)
0.242 0.319
0.058 0.014
-4 -3
-4 3 4
-1.96 -0.67 2.58
-2.58
50% of area 68.3% of area
95% of area
0.67 1.96
1 2 -2 -1
GRAPH OF THE NORMAL DISTRIBUTION FUNCTION
N(x)
0.2
-3 -2 -1 0 0.25 0.67 1 2 3 4
-4
0.4 0.6 0.7 0.75 0.8 0.9 0.95
1.0
0.52 0.84 1.28 1.96
0.5