BCGS2-M11

Dr. Najat MORADI FSTH

Descriptive statistics & probabilities

Page 1
 Descriptive statistics & probabilities

Table of contents

Chapter 1: Statistics with a Single Variable.............................................................................. 3

Chapter 2: Statistics with Two Variables..................................................................... 12

Chapter 3: Combinatorial analysis and the notion of probabilities............................... 20

Chapter 4: Discrete Random Variables................................................................... 29

Chapter 5: Continuous random variables……………………………………………… 39

References48

Page 2
 Descriptive statistics & probabilities

Chapter 1
Statistics with one variable

[Link]:
a-The word " Statisticsin the plural refers to collections of figures, presented
sometimes in the form of tables, sometimes translated by graphs, and which
gather all the observations made on numerous facts, andrelated to a
same phenomenon.
b-The " Descriptive statisticsis a set of methods for describing
and to analyze, systematically and in a quantified mannerstatistical data
c- The Mathematical statisticsis the science of which The purpose is to formulate laws
of behavior based on observation.
It relies not only on descriptive statistics, but also on calculation.
of probabilities.

[Link]:
The populationstatistics
is the set on which the study is focused.
Individualsor the statistical units
are the elements that make up the populationstatistics.
The sample
is a part of the population.
The character
is the aspect that we observe on individuals, it is the fact on which the study is focused
statistics.
He isit is often possible to observe two or more characters in the same
study.
When thecharacter is not not subject to a measure (for example nationality
of a person, their family situation, etc.), it is saidqualitativeand we call him/heralso
modalities.
When, on the contrary, this characteristic can be measured (for examplethe age weight, the
size.. etc), itis saidquantitativeand we call its measurevalueorvariable
statistics.
The statistical variablemaybe a name, and in this case it is saiddiscontinue
ordiscreet, as it can be an interval, and in this case it is calledcontinue.
Effective or absolute frequency

Page 3
 Descriptive statistics & probabilities

is the number of individuals having a value or a modality

The relative frequency
is equal to the absolute frequency divided by the total number of occurrences

2.1. Example 1
According toa
study conducted on a sample of 100 families, the number of children
aged 0 to 16 years is detailed in the following table:

The populationis the set of families

The characteris the number of children, it is a statistical variablediscreet
The total workforceis 100

2.2. Example 2
Distribution of deposits made in a bank, according to the amount in
thousands of dirhams:

The charactereastquantitativein the form of intervals , so it's a variable

statisticcontinue.
The total workforceis 500, andrelative frequenciesare as follows:

Page 4
 Descriptive statistics & probabilities

2.3. Example 3
In the study ofnumber of registrations, in thousands, of new vehicles in
France in 2007 for four major manufacturers, we have the following results:

The characterStudying here is the type of the new ones. vehicles, so it's a character
qualitative.

2.4. Example 4
Here we will give an example of a two-way statistical table, that is to say
a statistical study with two variables.
It is a distribution table of the children in a school based on their ages and their
weight.

Page 5
 Descriptive statistics & probabilities

3.1. Bar chart:

This bar chart is used to represent [Link] a painting
discrete quantitative statistical characteristic (discrete statistical variable).
On Cartesian coordinates, let's plot the values of the variable on the x-axis.
and on the ordinate the associated populationsor the frequencies

3.2. Histogram
To represent the counts or frequencies of a continuous statistical variable
or of a modality, we use the representation by histogram (expls2 and 3).
We carry on the abscissaabout at the end, segments that have the same amplitude,
representing the classes.
On each of the base segments, a rectangle is [Link] height the staff
partner.

Note:
In a histogram representation, it is the area of the rectangles that is
proportional tostaff.

Application to example 2

Page 6
 Descriptive statistics & probabilities

120

100

80

60

40

20

3.3. Polygon:
In a histogram, by connecting the midpoints of the tops of its [Link]
obtains a polygon that serves to give the general appearance of the phenomenon studied.

Application to Example 2

Polygon of staff
120
100
80
60
40
20
0

Page 7
 Descriptive statistics & probabilities

3.4. Curve:
We representthe increasing cumulative totals and the totals decreasing cumulative
by curves( de even for the cumulative increasing frequenciesand the
decreasing cumulative frequencies.

Application to the example1

The table of cumulative increasing frequencies for example 1 is as follows:

In the case of acontinuous statistical variable, we plot the frequency curve

growing cumulationsby associating the zero image with the minimum value of the first
interval, thus the maximum value of the last interval will correspond to the frequency
total.

Page 8
 Descriptive statistics & probabilities

Application to example 2

Quantitative statistical variables can be summarized by:

a) Of thecharacteristics referred to as "..." positionor of 'trend' centralwho

give importance tocore values
b) Of theso-called ‘ parameters’ dispersionwhich characterize the dispersion or the
concentration of values around theaverage.
We consider in all that follows the following statistical table:

4.1. Position parameters:

Let n be the total number i.e. n∑
=

Page 9
 Descriptive statistics & probabilities

[Link] averages:
a- The averagearithmetic is
∑ ∗
̅= =

∑ ∗
̅=
b- The quadratic meanis

∑ x *n
q=
n
c- The geometric meanest

g = ∏x

d- The averageharmonic is

ℎ=
∑
.

Remarks:
a-The average quadratic isthesquare rootof the arithmetic mean of x squared
b-The harmonic mean isthe inverseof the arithmetic meanof .
c-For a continuous statistical variable, the previous averages of
the same way by takingthe environmentsintervals as being the values

4.1.2. The median:

The median is the value that divides the statistical variable into [Link] to
equal staff.
In thefirst series, we find the values below the median.
In the second series, we find the values above the median.
It is generally noted by M.
To determine the Median is obtained using the table of cumulative frequencies. :
For a discrete statistical variable, the median is the value that has the
cumulative growing effective first strictly greater than
For a continuous statistical variable, the interval or classassociated with
first cumulative effect strictly greater than is saidthe class
medianand the median is a value of thisinterval. Its determination is

Page 10
 Descriptive statistics & probabilities

more practicalusing the curve of cumulative increasing frequencies (or also the
cumulative frequency curvedecreasing) since it is the antecedent of the halfof
the total workforce.

4.2. Dispersion or concentration parameters:

[Link] extent:
It is the difference between the smallest and the largest value.

[Link] and standard deviation:

The variance is: V = q² – m² whereWhat is the average? quadratic andm is there
arithmetic mean.
The standard deviation is: σ =

Usefulness of the standard deviation:

According to a result by the MathematicianBienaymé Chebyshevwhateverlet it be

statistical variablestudied, the intervalm- 2 ; m + 2 ] containsat least
75%of the studied population.

[Link] interquartile range:

The definition of this dispersion characteristic is based on knowledge.
of the concept of quartiles.
On appelle First Quartileand we refer to it as the value of the variable such
that 25% of the values taken by the variable are lower than it and 75%
superior.
We callThird Quartileand we designate by the value of the variable such
that 75% of the values taken by the variable are below it and 25%
superior.
As a result of thesedefinitions theSecond Quartile will blend inwith the
median.
The interquartile rangeis the difference - ).
For a V.S.D is the value having the first cumulative frequency increasing
strictly greater than .
And is the value having the first cumulative frequency greater than
strictly at .
For a V.S.C we calculate and Pu as for the median;by determining the
respective backgrounds of and in the curve of cumulative totals
croissants.
Page 11
 Descriptive statistics & probabilities

Chapter 2
Two-variable statistics

A statistic with two variables is a study of two different characteristics.

on the same population, as in example 4:the distribution of children
from a school according to their ages and their weights.

In fact, the data of astatistical series two variables are generally

represented initially as pairs ( variable1; variable2), and then
we group them together in the formof classes.

Example 5:
We have size and weight measurements of 19 adolescents presented by
pairs (thesize; the weight).

140; 38.2; {161 ; 44,3} ; {155 ; 46,1} ; {148 ; 38,2} ; {155 ; 50.5} ; {123 ; 22,4} ; {160 ;
40.4{140 ; 34.7} 165; 50.5{172 ; 50.5{155 ; 38.1} ; {160 ; 57.3} ; {142 ; 39.3
; {157 ; 46,1} ; {142; 37.1}; {148 ; 45,9}; {180 ; 66,3}; {167 ; 60} ;{165 ; 50,5}}

Page 12
 Descriptive statistics & probabilities

This data can be presented [Link] points of

coordinates (variable1; variable2) that forma point cloud

We can also group this data like this:

[20 ;40[[40; 60[ [60 ;80]Total

[120;140[ 1 0 0 1

[140 ;160[ 6 4 0 10

[160; 180] 0 6 2 8

Total 7 10 2 19

Page 13
 Descriptive statistics & probabilities

Graphiquement ce tableau est illustré by a three-dimensional histogram.

We consider the statistical table with two variables X and [Link]:

Page 14
 Descriptive statistics & probabilities

. = , ∊ {1, … , }

. = , ∊ {1, … , p}

= .= .

[Link] saidthe marginal effects of Y»

and theNl. are saidthe marginal staff of X

Marginal averagesof X and of Y are:

1
= Nn. * Xl

1
= . m * Ym

Marginal variancesare:

1
= Nl. * Xl −

1
= N. m * Ym² - Y²

The marginal standard deviationsare:

σ =

σ =

We denote by Cov(X,Y) et hyo ( , ) the following values:

( , )= ( ∗ )− ∗

Page 15
 Descriptive statistics & probabilities

( , )= ( , )

4.1. Graphic adjustment:

We will study, through a graph, whether there is a relationship between these two.
variables and what is the nature of this relationship.

We draw a curve that we callcurve of regression» and who passes between the
observations of a point cloud.
Most often, we try to draw aline that we then refer to as right
of regression» orright of trend».

And we will say that we haveone linear liaison» between the variables.

Page 16
 Descriptive statistics & probabilities

4.2. The analytical adjustment:

Analytical adjustment consists of determiningthe function whose curve
adapts in the most satisfactory way to the observations made, that is to say the
curve closest to the points of the cloud.

For this, we are looking for the function f such that the differences (− ( )
) with
∊ {1, … , p} andm ∊ {1, … , } , be minimal.

Page 17
 Descriptive statistics & probabilities

Indeed,the method of least squares consists of determining f such that

sum of the squares of these differences

∗ ( − ( )²

be minimal.

4.3. Linear adjustment:

For a linear adjustment, two parameters are determined.a and bsuch as the
line with equation y= . + , be close to the points of the cloud.

By the method of least squares, we find that:

( , )
=

= −

Remark

The regression line passes through the point with coordinates( ; ) that we call
midpoint.

Page 18
 Descriptive statistics & probabilities

4.4. Correlation coefficient:

Here we define another parameter that we callcoefficient of correlation» who
measure the more or less great dependence between the two characters X and Y.
It is denoted by the letter 'r', it varies between -1 and +1 and it is calculated as follows:

( , )
=
∗

Plus r is close to+1or of-1, plusthe two characters aredependents.

The closer he is to0, plus the two characters are independent.

Page 19
 Descriptive statistics & probabilities

Chapter 3
Combinatorial analysis and notion of
probabilities

1.1. Fundamental principle of analysiscombinatorics

If any procedure can be represented by n different ways, if
after this procedure, a second can be representedof n different ways
and if then a third procedure can be represented by n ways
different and so on

So the number of different ways to execute the procedures in

the indicated order is equal to the product

. . …

Example
Suppose a license plate contains two distinct letters followed by
of three digits, the first of which is different from zero How many plates
What different things can we print?

There are 26 different ways to print the first letter, 25 different ways
to print the second letter, 9 different ways to print the first digit and
10 for the other two.

It follows that one can print26 ⨯ 25 ⨯ 9 ⨯ 10 ⨯ 10 = 585000plaques

different

1.2. Arrangements andpermutations:

A permutation is an arrangement of n objects considered at the same time.
and taken in a given order.

The arrangement of p of these n objects (withp ≤ n) in a given order is said

arrangement of index p of n objects.

Page 20
 Descriptive statistics & probabilities

We note by A the number of thesearrangement of index p

Theorem 1:
!
=
( − )!

With n! = 1 × 2 × ... × n is the product of the positive integers from 1 to n

Read "factorial n ».

And by convention 0!= 1.

If p=n, = !, Therefore, the number of permutations of n objects is n!

1.3. Permutations with repetition:

Theorem 2:
The number of permutations of n objects whose n are similar, n are
similar ... and are similar is

!
!⨯ !⨯…⨯ !

1.4. Combinations:
A combination of p objects from n is a subset containing p.
elements.

We designate by ou the number of combinations of p objects out of n.

Theorem 3:

= =
!

Properties:

C =1=
C = =
3. C = for everything1 ≤ p ≤ n.

Page 21
 Descriptive statistics & probabilities

4. C + = for everything1≤p≤n

5. Thebinomial formula:
For all a and x in R and all n in N we have: + (= ) ∑

1.5. Tree diagram:

A tree diagram is a convenient way to count all the
possible outcomes of a series of experiments each of which can occur once
finite number of times.

Example:
Determine the elements of the product set AxBxC with A={1,2}, B={a, b, c}

and C={3,4

Page 22
 Descriptive statistics & probabilities

2.1. Ensemblefundamental and events :

2.1.1. Definitions:
The set S of all possible outcomes of a given experiment is called
fundamental ensemble.

Each result, that is, an element of S is called a point of S.

An event A is a subset of the sample space S.

The event {a} consisting of a single point a of S is called an event

elementary.

The empty set ∅ is the impossible event.

S is the sure or certain event.

A ∪ Bis the event that occurs if A or B is realized.

A ∩ Bis the event that occurs if both A and B are realized.

̅
The complement of A, denoted by , is the event that occurs if A
is not accomplished.

2.1.2. Examples:
Example1:
We roll a die and observe the result obtained.

The fundamental set isS { 1, 2, 3, 4, 5, 6 }

Let A, B, and C be the events corresponding respectively to the occurrence of a

even number, an odd number and a prime number.

A = { 2, 4, 6 }B = { 1, 3, 5 } and C{ 2, 3, 5 }.

A ⋂ B = ∅,A ∪ C = { 2, 3, 4, 5, 6 } and B⋂ C = { 3, 5 }.

Page 23
 Descriptive Statistics & Probabilities

Example2:
We toss a coin and observe the sequence of heads H and tails T.
obtained.

The fundamental set is: S = {FFF, ["FFP","FPF","FPP","PFF","PFP","PPF","PPP"] }.

Example3:
We toss a coin until we get heads, and we count the
number of times the coin has been tossed.

The fundamental set is: S = { 1, 2, 3, . . . , ∞ }.

This is an example of an infinitely countable fundamental set.

2.2. Axiomsof probability calculations :

Let S be a fundamental set, E the family of events, and P a function to
real values defined on E.

Definition:
It is said that P is a probability function and that P(A) is the probability of
event A if we have the following axioms:

[Link] each event A 0<= P A ≤1. ( )

II. P(S)= 1.

III. If A and B are mutually exclusive events then,

P(A ∪ B=) ( +)( ) .

[Link] A , A , A ... is a sequence of events that mutually exclude each other.

so, P A ( Union A Union A ∪. . . P) A + ( (A
) ( +) P) A + . . .

Consequences:
1) P( ∅)equals zero.

2)P A( )= 1 - P(A)with A is the complement of A.

Page 24
 Descriptive statistics & probabilities

3)For two events A and B a :

P (AB = ) P(A)- P(A ∩ B) and
P (A∪ B = ) P A+( )P B -(P(A∩
) B)

4) A, B, and C being events. We have :

P (A∪ B ∪ C )
= P (A+) P B( +) P C-(P )A ∩ (B- P A )∩ C-( P C ∩) B+ P(A
( ∩ B) ∩ C)

2.3. Ensemblesfinished probabilities :

Let S= { a , a , a , . . , a a finite fundamental set.

Definition:
If at each point a to S a real number p is attributed , which is called
probability of a checking the following properties:

I. For all k we have 0≤ p ≤ 1.

II. The sum of p is equal to 1: p + p + . . . + p= 1.

It is said that S is a finitely probabilized ensemble.

We then define the probability P(A) of any event A as being the

sum of the probabilities of points A.

Example4 :
We toss three coins into the air and observe the number of heads obtained.

The possible outcomes are:

PPP;PPP PFP ;PFF ; FPP ;FPF ; FFP ;FFF}.

But the fundamental set is S= {0 ; 1 ;2 3 }.

We ask: p 0( =) p; 1 = (p; )2 = ( ) and (p3) = .

S is thus a finite probabilistic set.

Let A be the event that we have at least one head and B be the event that
would correspond to either three heads or three tails, that is to say thatA =2 3} and
B = {0 ;3}.

Page 25
 Descriptive statistics & probabilities

So we have P( A=
) 1 +(p )2 + p( 3=
) ( ) and (P )B= p (0 )+p 3 (= )

2.4. Ensemblesequally probable outcomes:

A finite probabilistic set S, where each element has the same probability is
called equiprobable or uniform space.

If S contains n points then the probability of each point is .

On the other hand, if an event A is composed of r points, its probability is then

r× = .

Example 5:
When throwing a fair die.

2) When drawing a card from a standard deck of cards.

3) When three coins are tossed wellbalanced.

3. 1. Probabilityconditional:
Let E be an arbitrary event from a fundamental set S such that P(E) 0.

Definition:
The probability of event A occurring, given that event E has occurred.
previously, or in other words, the conditional probability of A given that
( ∩ )
It is realized that we write (A) is defined by: ( =
)
( )

Example6 :
We throw a pair of fair dice.

Knowing that the sum is equal to 6, calculate the probability that at least one
some dice had given a 2.

For this we have:

E = The sum equals 6

Page 26
 Descriptive statistics & probabilities

= { (1 ; 5) , (2 ; 4) (3 ; 3), (4 ; 2), (5 ; 1)}

A = 'Auless one of the dice gives 2

= { (2 ; 1), (2 ; 2), (2 ; 3), (2 ; 4), (2 ; 5), (2 ; 6), (1 ; 2), (3 ; 2), (4 ; 2), (5 ; 2), (6 ; 2)}
( ∩ )
A ∩ E = { (2 ; 4), (4 ; 2) }.Where from ( ) = =
( )

Multiplication Theorem:
A and E two events such that: P(E)0. So we have P A∩ =( ). ( )( )

This theorem can be generalized for any events.

A , A , A . . . A and we prove that we havethe result following:

P(A ∩ A ∩ A ∩ . . .∩ A )

= (A ). A( . ) ∩ A( . ( ) )∩ ∩ A . . . . ∩ ∩ ∩ . . .∩ (A)

Bayes' Theorem:
Let B be aeventandA , A A . . . A a partition of S, that is to say that:

The events A , A , A . . . A exclude each other mutually two by two and

A ∪ A Union A ∪ . . .∪ A = .

We then have the following equality:

P( B=
)( ) A( . ) B ( +)P A (. ) B +P
( )A . ( B) +. . ( .+) A . ( )
B

Example7 :

We toss a biased coin in such a way that: p(F) = and p(P) =

If it is heads that appears, we randomly choose one of the numbers from 1 to 9.

If it isstack that we obtain, we choose randomly one of thenumbers ranging from 1 at 5.

Calculate the probability p that an even number was chosen.

We consider the following events:

Page 27
 Descriptive statistics & probabilities

To haveface To havepile and P = ‘’ Having an even number

So then,

p = P(P′)

= PP( ∩ F )+P P ( ∩ P )

= (F )(. ) P ′ + P ( )(P′ )

= × + × =

It is the theoremof Bayes

One can read this requested probabilityon the tree diagram with the
corresponding probabilitiesas follows :

Note that there are two paths leadingto P' and the probability of this
event isthe sum of the products of probabilities mentioned all over a
same path.

3. 2. Independence:
Definition:
It is said that an event B is independent of an event A with P(A) ≠0 yes :

(B) = ( )

Consequence:
A and B are independent if and only if P(A∩ B) = P(A).P(B)

Page 28
 Descriptive statistics & probabilities

Chapter 4
Discrete random variables

1.1 Definition
A random variable X on a fundamental set S is a function of S
in the set ℝ of real numbers.

If S is a discrete set, any real-valued function defined on S is a

random variable. And in this case, we are talking about adiscrete random variable.

1.2 Example
We toss a coin three times, and observe the sequence of heads (H) and tails (T)
of faces (F) obtained.

Consequently, the fundamental set is:

={ ; ; ; ; ; ; ; }

Let X be the random variable thateach point of S corresponds to the number

fromfaces who appeared. We have:

X(PPP) = 0,X(FPP) = X(PFP) = X(FPP) = 1, X(FFP) = X(FPF) = X(PFF) = 2

and X(FFF)= 3.

The image space of X is X(S){0, 1, 2, 3}

1.3 Operationson random variables :

We consider two random variables.X and Y defined on the same set

fundamental S.

Page 29
 Descriptive statistics & probabilities

Property:
Random variables are definedX+Y, X + k,k.X and X.Y ( where k is a number
real) on the ensemble fundamental S in the following way:

For every s in S we have:

(X + Y)(s)= X(s) + y(s),(X +k)(s) = X(s) + k,

(k.X)(s) = k.(X(s)) (X.Y)(s) = X(s).Y(s).

Notations :
We use abbreviated notations = ( )( ≤ ≤ ) to designate the
probabilities :

( = )= ({ ∊ ( ) = }

( ≤ ≤ )= ({ ∊ (≤ ) ≤ }

2.1 Definition:
Let X be a random variable on a finite value sample space S. It is
to say ( )= { , , . . ., }.

X(S) becomes a probabilistic space if we define the probability P(X = ) of the

that will write it f( ).

This function f on X(S) which is defined byf( )= P( X = is it that we

call the distribution or the probability law ofX, which is usually given
in the form of a table:

… …
( ) ( ) ( ) … …

The lawThe probability f satisfies the following conditions

( For all i,f( x ) greater)than or equal to( 0and

) fx =1

Page 30
 Descriptive statistics & probabilities

2.2 Definition ofthe mathematical expectation :

If Xis a random variable with the previous probability law, the mean
or hopemathematics of the variable X denoted by E(X) is defined
by:

( )= × ( )

Page 31
Descriptive statistics & probabilities

Page 32
Descriptive statistics & probabilities

Page 33
Descriptive statistics & probabilities

Page 34
Descriptive statistics & probabilities

Page 35
Descriptive statistics & probabilities

Page 36
Descriptive statistics & probabilities

Page 37
Descriptive statistics & probabilities

Page 38
Descriptive statistics & probabilities

Page 39
Descriptive statistics & probabilities

Page 40
Descriptive statistics & probabilities

Page 41
Descriptive statistics & probabilities

Page 42
Descriptive statistics & probabilities

Page 43
 Descriptive statistics & probabilities

Normal approximation of the binomial distribution:

The distributionbinomial is very close to the distributionnormal, provided that n

be great and neitherp, ni that are nottoo close to zero.

Page 44
 Descriptive statistics & probabilities

Example:
X follows the binomial law B(n = 8 ; p = 0.5) don'tthe probability law is given by the
next table:

Graphically, it is noticeable that thisdistribution can be approached by a

normal distribution.

Comparative representations of binomial and normal distributions.

The previous property of the normal probability law is generalized thanks to the
Central Limit Theoremfollowing:

Central Limit Theorem:

Let (X ) a sequence of random variables defined on the same space
probabilized, two by twoindependent, and all under the same law of expectation
mathematics ofvariance σ .

We note S = ∑ the sum of random variables ; ; . . .(; and )

= ∑ their average.

Page 45
 Descriptive statistics & probabilities

So:
µ
( )=µ, ( )= and for sufficiently large n, the law of = ; the
√
standardized variable associated with M is adjacent to the centered normal law
reduced N(0;1).

Example:
We throw180 times a well-balanced die.

Calculate the probability that face 6 appears between 29 and 32 [Link]

includes.

Let X be the number of times wherethe

face 6 appears, then X follows the binomial law B(180;
p = 1/6 and E(X) = n.p = 30 andV(X) = n.p.q = 25.

To calculate P(29 ≤ X ≤ 32), on apply the Gaussian [Link] is to say

what X≈ Y with Y follows the normal law N(30; 5) of such sort of E(X)= E(Y) = 30
and V(X)= V(Y) = 25.

So by transitioning from discrete to continuous, wewrite:

P ( 29≤ X ≤ 32≅ ) 29 (- 0.5 ≤Y ≤ 32 + 0.5 )

Therefore
∗
P ( 29≤ X ≤ 32≅ ) 28.5( ≤ Y ≤ 32.5 = P -0.3
) ≤( ≤ 0.5 = 0.3094)

Page 46
Descriptive statistics & probabilities

Page 47
 Descriptive statistics & probabilities

1- Fabrice MAZEROLLE
Descriptive statistics
Gualino publisher(2006)

2-N. BRILLOUET – BELLUOT

Introduction to probability and statistics
[Link]

3-Seymour LIPSCHUTZ
Probability
Schaum Series

[Link]
Essential notions of Statistics and probability calculation.
Sirey 1970

Page 48