Topic 2B. Modelling With Functions

International Baccalaureate
MATHEMATICS
Applications and Interpretation SL (and HL)
Lecture Notes
Christos Nikolaidis
TOPIC 2
FUNCTIONS
2B. Modelling with functions
2.13 LINEAR MODEL ..……………………….…………………………………………………………… 1
2.14 QUADRATIC AND CUBIC MODELS …………………………………………….………… 6
2.15 POWER FUNCTION MODEL – DIRECT/INVERSE VARIATION ……….….. 13
2.16 EXPONENTIAL MODEL ..………………………………………………………………………… 20
2.17 SINUSOIDAL MODEL .………….……………………..….……….……………………….…..…. 30
2.18 AN APPLICATION OF LINES – VORONOI DIAGRAMS ………………..………. 36
2.19 PERCENTAGE CHANGE – FINANCIAL APPLICATIONS ………………………. 40
Only for HL
2.20 LOGISTIC MODEL ……………………………………………………………………………………. 53
2.21 NATURAL LOGARITHMIC MODEL …………………………………………………….…... 56
2.22 LINEARISATION OF THREE MODELS …………………………………………………… 59
2.23 PIECEWISE MODEL ………………………………………………………………………………… 63
June 2022
TOPIC 2: FUNCTIONS Christos Nikolaidis
2.13 LINEAR MODEL
 CONNECTION WITH ARITHMETIC SEQUENCES

Consider the arithmetic sequence with first term u1  9 and
common difference d  3
9, 12, 15, 18, 21, 24, …
Let us represent the terms of this sequence in the Cartesian plane.
We observe that the points lie on a line. This is not an accident.

The n-th term of the sequence is given by
un  9  (n- 1)  3
that is
u n  3n  6
which is a linear expression. In fact the points lie on the line
y  3x  6
The gradient m of the line is equal to the common difference d=3.
1
 A POINT SERIES ON A LINE
Consider the following data
x 1 2 5 7 10 12
y 6.1 7.2 10.5 12.7 16 18.2
Let us place the points (x,y) on the Cartesian plane
It seems that the points lie on a line. Indeed,
The gradient of the line segment AB is

Δy 1.1
m AB    1.1
Δx 1
The gradient of the line segment BC is

Δy 3.3
m BC    1.1
Δx 3
Even for non-consecutive points, for example C and F,

Δy 7.7
m CF    1.1
Δx 7
In fact, any pair of points gives the same gradient, and this is the
gradient of the line in question.
What is the equation of this line?
2
Two points are enough to find the line passing through them.
Let us select the endpoints A(1,6.1) and F(12,18.2).
First approach
The gradient of the line segment AF is

Δy 12.1
m AF    1.1
Δx 11
Thus, the equation of the line (in point-gradient form) is
y-y1 = m(x-x1)
 y-6.1 = 1.1(x-1)
We carry on finding the gradient-intercept form:
y-6.1 = 1.1x-1.1
 y = 1.1x+5
Second approach
The equation of the line has the form

y = ax+b
We need to find the coefficients a and b:
By substituting A(1,6.1) we obtain a+b=6.1
By substituting F(12,18.2) we obtain 12a+b=18.2
The solution of the system is a=1.1 and b=5.
Therefore, the line is
y = 1.1x+5
Notice: any pair of points would give the same result (check!)
Confirm
If we pick another point among our data, say E(10,16), it satisfies
the equation: For x=10,
10+5 = 11+5 = 16
1.1x+5 = 1.1
3
EXAMPLE 1
The relation between Celsius degrees (°C) and Fahrenheit (°F) is
linear. That is
°F = a °C+ b
We are told that
4°C correspond to 39.2 °F
10°C correspond to 50 °F
Thus
39.2 = 4a+b
50 = 10a+b
The solution is a=1.8 and b=32. Therefore the formula is
°F = 1.8 °C+ 32
In other words in order to convert temperatures in degrees Celsius
to Fahrenheit, multiply by 1.8 (or 9/5) and add 32.
 A POINT SERIES ALMOST ON A LINE
Sometimes in real life the data seem to almost lie on a line. For
example,
Q 10 13 15 20 25
P 15 20 25 33 45
4
A first thought to model this linear relation could be to pick two

representative points, say (10,15) and (25,45).
If
P = aQ +b
then
(10,15) gives 10a+b=15
(25,45) gives 25a+b=45
The solution is a=2, b=-5, hence
P = 2Q - 5
If we add the line y=2x-5 on the graph above we obtain
Our model seems to fit quite well to the given points.
 REGRESSION – BEST FIT LINE BY USING GDC
Later, in Statistics (Topic 4), we will discuss in further detain the

method of regression. At the moment we simply present the best
fit line that our GDC gives. For our data above:
MENU: Statistics: Enter List 1: Q-values List 2: P-values
CALC: REG: X [F1]: ax+b [F1]
The best fit line is

P = 1.98Q - 5.22
5
2.14 QUADRATIC AND CUBIC MODELS
 POLYNOMIALS AND GRAPHS
Remember the following functions
Linear function y=ax+b

Quadratic function y=ax2+bx+c
We also say called them polynomials of degree 1 and 2 respectively.

The degree is the largest exponent of x in this expression (all the
exponents are positive integers; we may have a constant term as
well.
Along the same lines we define the polynomial of degree 3 which

also called
Cubic function y=ax3+bx2+cx+d
The simplest polynomials up to degree 3 are the following
Line quadratic cubic

y=x y=x2 y=x3
y=-x y=-x2 y=-x3
6
For any linear function y=ax+b with a≠0

 There are 2 parameters a and b; so we need two points to
determine the values of a and b (by solving a system).
 The graph is a straight line
increasing if a>0
decreasing if a<0
For any quadratic function y=ax2+bx+c with a≠0

 There are 3 parameters a, b, c; so we need three points to
determine the values of a, b and c.
 The graph is a parabola
concave up if a>0
concave down if a<0
For any cubic function y=ax3+bx2+cx+d with a≠0

 There are 4 parameters a, b, c, d; so we need four points
to determine the values of a, b, c and d.
 The graph has an S-shape.
it looks like if a>0
it looks like if a<0
For example, the graph of the cubic function

y  x 3  6x 2  11x  6
Is
Notice: As x→+∞ (going to the right), y tends to +∞, since a>0.
7
 QUADRATIC MODEL
If they give us three points on a parabola y=ax2+bx+c
Points : (1,-3) (3,-1) and (4,6)
we can find the parameters a, b and c by substituting the

coordinate of the points:
(1,-3) gives a+b+c = -3
(3,-1) gives 9a+3b+c = -1
(4,6) gives 16a+4b+c = 6
We solve the system by GDC: a=2, b=-7, c=2
Hence, the quadratic function is
y=2x2-7x+2
If we have a number of points which form a U-shape with certain

properties, for example
 There seems to exist a vertical axis of symmetry

 There is a vertex
 The points increase or decrease in a “smooth” way
which is not exponential
we may assume that the U-shape is a parabola. This will be our

model. Of course we are not certain that the assumption is correct
but it is a practical approach if we cannot find an analytic solution.
8
EXAMPLE 1
The perimeter of a rectangular region is 100m. What is the area?

Is it always the same? If not, what is the maximum possible area?
A model approach
Suppose that the length of one side is x. Choose some values for x:
one side (x) the other side Perimeter Area (A)

5 45 100 225
10 40 100 400
20 30 100 600
25 25 100 625
30 20 100 600
37 13 100 481
Let us observe the points (x,A) on the Cartesian plane
x 5 10 20 25 30 37
A 225 400 600 625 600 481
It seems that the points lie on a parabola with axis of symmetry

x=25. Let us pick 3 representative points,
the first point (5,225)

the last point (37,481)
and the point (25,625) which seems to be the vertex.
9
If A=ax2+bx+c then
point (5,225) gives: 5 2a + 5b + c = 225
point (37,481) gives: 372a + 37b + c = 481
point (25,625) gives: 252a + 25b + c = 625
The GDC gives the solution a=-1, b=50, c=0.
That is
A=-x2+50x
Indeed, the remaining points satisfy the equation, for example let
us check the point (10,400):
A=-102+5010 = 400
The analytic approach

Let x be the length of one side and y the length of the other side.
Then
Perimeter = 100
 2x+2y = 100
 x+y = 50
 y = 50-x
Hence, the area is

A = xy = x(50-x)
i.e. A = 50x–x2
The maximum possible area is obtained at x=25 (vertex), so
Amax = 500 m2
Notice: Here the analytic approach is easy. But this is not always
the case and sometimes the model approach is our only option.
We chose this example to demonstrate the two different
approaches.
10
In general, for the quadratic
f(x) =-x2+50x the domain is xR.
But for our particular problem x can be between 0 and 50. So our
model is
A=-x2+50x, 0<x<50
 A POINT SERIES ALMOST ON A PARABOLA - REGRESSION
If the points do not perfectly fit into a parabola, we can pick three
representative points and find a quadratic model that describes
quite well our data.
Again, a better solution could be found by using regression.

For our example,
x 5 10 20 25 30 37
A 225 400 600 625 600 481
the points, of course, perfectly fit into a parabola, so the GDC will
give the exact solution. But let us try it just for practice:
MENU: Statistics: Enter List 1: x-values List 2: A-values
CALC: REG: X2 [F3].
The (best fit) quadratic model is

A=-x2+50x
11
 CUBIC MODEL - REGRESSION

Similarly, if we observe that our data form an S-shape
we can assume that our model is a cubic function of the form
y=ax3+bx2+cx+d
If we know that the points fit exactly on a cubic function:
We pick 4 points and substitute on the model to find a, b, c, d

(by solving a system of 4 equation).
If the points almost fit on a cubic function:
A less accurate model is found by picking 4 random points and

work as above.
A better model can be obtained by using regression
MENU: Statistics: Enter List 1: x-values List 2: y-values
CALC: REG: X3 [F4].
12
2.15 POWER FUNCTION MODEL – DIRECT/INVERSE VARIATION
Let us concentrate on the functions of the simple form

y  xn
We have already seen the graphs of the functions y  x 2 , y  x 3 but

less us investigate more powers (i.e. values of n).
Use your GDC to verify all the results below:
when n is positive when n is negative
(n  1) ( n  1 )
1
yx y
x
( n  2 )
(n  2)
1
y  x2 y
x2
(n  3) ( n  3 )
1
y  x3 y
x3
(n  4) ( n  4 )
1
y  x4 y
x4
y-axis is a vertical asympote
13
We observe that when n is even, i.e. 2, 4, 6, …
the graph is symmetric about y-axis
or
when n is odd, i.e. 1, 3, 5, …
the graph is symmetric about the origin (0,0)
or
 POWER MODELS
In real life applications we meet models of the form
k
y  kx n or y n
x
where x is takes only positive values.
For positive exponents the models look like as follows
14
For negative exponents the models look as follows
 DIRECT AND INVERSE VARIATION
The relation
y  kx n
shows a direct variation between (a power of) x and y:
when x increases then y increases as well

Notice:
If y = kx we say y is (directly) proportional to x
If y = kx 2 we say y is (directly) proportional to the square of x
The relation
k
y
xn
shows an inverse variation between (a power of) x and y:
when x increases then y decreases

Notice:
k
If y = we say y is inversely proportional to x
x
k
If y = we say y is inversely proportional to the square of x
x2
15
EXAMPLE 1
Consider the relation
P = 5Q 2
Q 1 2 3 4 …
P 5 20 45 80 …
The quantity P is proportional to the square of Q.
EXAMPLE 2
Consider the relation
5
P =
Q2
Q 1 2 3 4 …
5 5 5
P 5 …
4 9 16
The quantity P is inversely proportional to the square of Q.
EXAMPLE 3
(a) P is proportional to the cube of Q (i.e. P =kQ3).
Find the relation if we know that P = 5 when Q = 2.
(b) Suppose that P =kQn. Find the relation if we know that

P = 6.4 when Q = 2, P = 48.6 when Q = 3.
Solution
5
(a) 23k=5  8k = 5  k = = 0.625. Hence P = 0.625Q3
8
(b) 2nk=6.4 and 3nk=48.6. If we divide the two relations
n
3 n k 48.6 3
n
     7.59375 . Hence n = 5.
2 k 6.4 2
The first relation gives: 25k=6.4. Hence k = 0.2.
Therefore, P = 0.2Q5
16
Notice also that the relation

BC 2
A =k
D3
between the quantities A, B, C, D (k constant) implies that:
A is proportional to B
A is proportional to the square of C
A is inversely proportional to the cube of D
(when the remaining quantities stay constant).
 AREA vs LENGTH
Look at the area A for several shapes below
Shape Area
Square
A  x2
of side length x
x
Rectangle x
A  2x 2
of sides x and 2x
2x
Right-angled triangle 1 2
a A a
of equal vertical sides a 2
a
Circle r
A  πr 2
of radius r
In all these relations we observe that the area A is proportional

to the square of a length (side, radius, etc).
For a rectangle of sides x and y, since A =xy
 A is proportional to x (if y remains constant)

 A is proportional to y (if x remains constant)
 y is inversely proportional to x (if A remains constant). Why?
17
EXAMPLE 4
Consider all rectangles of area 50 with side lengths x and y
(a) Describe the relation between the x and y.
(b) Express the perimeter of the rectangle in terms of x.
Solution
50
(a) xy =50  y  . Thus y is inversely proportional to x.
x
100
(b) P = 2x+2y = 2x 
x
 VOLUME vs LENGTH
Look at the volume V for several shapes below
Shape Volumes
Cube of edge length x V  x3
Cuboid of square base of side x and height 2x V  2x 3

4
Sphere of radius r V πr 3
3
Cylinder with radius of base r and height also r V  πr 3

1
Cone with radius of base r and height also r V πr 3
3
In all these relations we observe that the area V is proportional

to the cube of a length (side, radius, etc).
1
For a cone of radius r and height h, since V  πr 2 h
3
 V is proportional to h (if r remains constant)
 V is proportional to the square of r (if h remains constant)
When V remains constant, say V = 100, then

1 300
πr 2 h  100  πr 2 h  300  h 
3 πr 2
Hence,
 h is inversely proportional to the square of radius r.
18
 A POINT SERIES ALMOST ON A POWER MODEL
Consider
x 1 2 3 4 …
y 2 15 50 120 …
Given that the points are well described by a power function model
of the form
y  ax n
the best power model can be obtained by using regression
CALC: REG: [F6]: Power [F3].
y  1.98x 2.95
19
2.16 EXPONENTIAL MODEL
 CONNECTION WITH GEOMETRIC SEQUENCES

Consider the geometric sequence with first term u 1  30 and
common ratio r  3
30, 90, 270, 810, …
Let us represent the terms of this sequence in the Cartesian plane.
The n-th term of the sequence is given by

3n
u n  30  3 n-1  30 
3
that is
u n  10  3 n
which is an exponential expression. In fact the points lie on the line
y  10  3 x
20
 THE FORM OF AN EXPONENTIAL MODEL
A simple exponential model may have the form

y  A bx if y is increasing
y  A  b-x If y is decreasing.
(assuming that b>1).
We have to determine two parameters, A and b.
Very often the model is given in the form:
y  Ae kx
According to k
y increases if k>0
y decreases if k<0
Still, we have to determine two parameters, A and k.
EXAMPLE 1
Suppose that the graph of an exponential function
y  A bx
passes through the points (1,30) and (4,810).
Then
point (1,30) gives: A  b 1  30 (1)
point (4,810) gives: A  b 4  810 (2)
If we divide (2) by (1) we obtain
A  b 4 810
  b 3  27  b  3
A  b1 30
Then (1) gives
A  3 1  30  A  10
Thus, the function is

y  10  3 x
21
Things are even better if we know the y-intercept of the function

(corresponding to x=0).
EXAMPLE 2
Suppose that the exponential function
y  Ae kx
has y-intercept (0,25) and passes through the point (5,100).
Then
point (0,25) gives: Ae 0  25  A  25
Now point (5,100) gives: Ae 5k  100
 25e 5k  100
 e 5k  4
We can use the natural logarithm ln (or directly GDC):
 5k  ln4
ln4
k  0.277
5
Hence
y  25e 0.277x
 EXPONENTIAL GROWTH OF A POPULATION
In many applications a quantity increases or decreases

exponentially according to time. For example:
The population P at time t (let’s say in years) is given by
P=Poekt
 If k>0, the population increases

 If k<0, the population decreases
For t=0, then P = Poe0 = Po
So the coefficient Po is always the initial value of P.
22
EXAMPLE 3
The number n of some particles at time t hours is given by
n=1000e0.2t
(a) What is the initial number of particles?
(b) What is the number of particles after 3 hours?
(c) The number of particles is 2500 after t hours. Find t.
(d) The number of particles doubles after t hours. Find t.
Solution
(a) For t=0, n=1000
(b) For t=3, n=1000e0.23=1822
(c) If n=2500 then 2500 = 1000e0.2t
GDC gives t = 4.58 hours.
Otherwise,
2500 = 1000e0.2t 2.5 = e0.2t
 ln2.5 = lne0.2t
 ln2.5 = 0.2t
ln2.5
t = = 4.58 hours
0.2
(d) It’s a similar question to (c). We let n=2000 and find t.
[even if we don’t know the initial value no or we let n=2n0]
2n0 = n0e0.2t  2 = e0.2t
 ln2 = lne0.2t
 ln2 = 0.2t
ln2
t = = 3.47 hours
0.2
We can also use the graph mode of GDC to answer the questions
23
Solving EXAMPLE 3 by using GDC-graph
The function
n=1000e0.2t
has the following graph
For question (a) we find the y-intecept: y=1000
For question (b) we use GSolv-Y calc: x=3 then y=1822
For questions (c) and (d) we use GSolv-X calc: y=2500 then x=4.58
y=2000 then x=3.47
Sometimes the constant k is not known. But we are given some

information to estimate it.
EXAMPLE 3
Suppose n=n0ekt. If the number of particles doubles every 4 hours

find the value of k.
For t=4, 2n0 = n0ek4  2 = e4k

 ln2 = ln e4k
 ln2 = 4k
ln2
 k = = 0.173
4
24
EXAMPLE 4
The mass m of a radio-active substance at time t hours is given by
m = 4e–kt
(a) The mass is 1 kg after 5 hours. Find k.
(b) What is the mass after 3 hours?
(c) The mass reduces to a half after t hours. Find t.
Solution
(a) For t=5, m=1, thus 1 = 4e–k5
GDC-SolveN gives: k = 0.277

m = 4e–0.277t
(b) For t=3,
m = 4e-0.2773 = 1.74
(c) For m=2, 2 = 4e–0.277t
Either by GDC-SolveN: t=2.50 hours
or by using ln:
 e–0.28t = 0.5  lne–0.277t = ln0.5  –0.277t = ln0.5

ln0.5
t = = 2.50 hours
- 0.277
This time is known as half-life time.
Again, we can answer the questions (b) and (c) above by using the
graph of m = 4e–0.277t
25
 THE EXPONENTIAL MODELS y  Ab x  c , y  Ae kx  c

If we add a constant number c to our simple exponential models
they take the form
y  Ab x  c if y increases
y  Ab -x  c if y decreases
or
y  Ae kx  c with k>o if y increases – k<0 if y decreases
We simply shift the graph c units above.
If the function is increasing it looks like
In either case:
The horizontal line y=c is a horizontal asymptote to the graph.
For x=0, the y-intercept is y = A+c
26
EXAMPLE 5
The population P after t years is given by
P  1000e -0.2t  500
The graph of this function is
The initial population is

P  1000e 0  500  1500
The population after 5 years is

P  1000e -1  500  867.87...  868
The population drops below 600 when P  600 .

Either by GDC-GSolv-X Calc: t = 11.51
Or by solving the equation
1000e -0.2t  500  600
 1000e -0.2t  100
 e-0.2t  0.1
 0.2t  ln0.1
ln0.1
t =11.51
 0.2
As t →+ the value of P approaches 500. It never drops below 500

(in mathematical terms the line P=500 is a horizontal asymptote)
27
EXAMPLE 6 (mainly for HL)
Consider the following table of data.
t 1 2 3
P 20 30 50
It is given that the points lie on the exponential function of the

form
P  Ab t  c
There are three parameters to be determined, namely A, b and c.
Let us substitute the coordinates of the given points:
The point (1,20) gives: Ab  c  20 (1)
The point (2,30) gives: Ab 2  c  30 (2)
The point (3,50) gives: Ab 3  c  50 (3)
We firstly eliminate c:
(2) – (1) gives Ab 2  Ab  10 (4)
(3) – (2) gives Ab 3  Ab 2  20 (5)
Now we divide (5) by (4) to obtain
Ab 3  Ab 2 20

Ab 2  Ab 10
Ab 2 (A  1)
 2 b2
Ab(A- 1)
Then (4) gives: 4A  2A  10  2A  10  A  5
Finally (1) gives: 10+c=20  c  10
Therefore, the model is
P  5  2 t  10
28
 A POINT SERIES ALMOST EXPONENTIAL - REGRESSION
Consider the example of the previous paragraph (where we found a

power function model.
x 1 2 3 4 …
y 2 15 50 120 …
Mind the difference between
Power model: y  ax n
bx
Exponential model: y  ab x (or y  ae )
2.95
The power function model we found was y  1.98x
Let us find an exponential model
either in the form y  aebx

or in the form y  ab x
The GDC, regression gives
CALC: REG: [F6]: EXP [F2]: [F1] or [F2]
[F1] gives
y  0.707e1.35x
[F2] gives
y  0.707  3.85 x
29
2.17 SINUSOIDAL MODEL
This paragraph presupposes the study of the trigonometric

numbers sine and cosine : see Topic 3 (Trigonometry).
For our purpose we only present some basic notions of the functions
f(x)=sinx, f(x)=cosx and certain modifications of them.
 f(x) = sinx°
Look at the graph of the function f(x) = sinx° on your GDC by using
the following settings
Set up: Angle: degrees
Set V-Window: x from -360° to 720°

y from -2 to 2
We obtain
The following table contains some values of the function which are
also shown on the graph above.
x 0° 90° 180° 270° 360° …
y=sinx° 0 1 0 -1 0
We have:
Domain: x R
Range: y  [-1,1] [since ymin=-1 and ymax=1]
We also observe that the curve is repeated every 360°
30
Furthermore, we define
Central axis: y = 0 (the average between ymax and ymin)

Amplitude = 1 (distance between ymax and central line)
Period: T= 360° (the length of a complete cycle)
 f(x) = cosx°
A similar curve is obtained for the function f(x) = cosx°
x 0° 90° 180° 270° 360° …
y=cosx° 1 0 -1 0 1
Again Domain: x R
Range: y  [-1,1] [ymin=-1 and ymax=1]
Central axis: y = 0
Amplitude = 1
Period: T= 360°
NOTICE:
 For both functions y=sinx and y=cosx, the horizontal distance
between two consecutive xmax = 360° (one period)

between two consecutive xmin = 360° (one period)
between consecutive xmax and xmin = 180° (half a period)
 The central axis is also known as principal axis.
31
We distinguish four basic types of trigonometric functions:
sinx -sinx cosx -cosx
y-intercept y-intercept y-intercept y-intercept

central/going up central/going down max min
Notice that the amplitude for all these functions is 1 (positive)
 SINGLE TRANSFORMATIONS OF sinx
Compare the original function f(x)=sinx (in black) with the following
three transformations of sinx (in red)
f(x)=sinx + 2 f(x)=2sinx f(x)=sin2x
The curve is shifted 2 The curve is stretched The curve is shrinked

untis up vertically (by 2) horizontally (by 2)
Characteristic that changed:
new new new

Central axis: y=2 Amplitude = 2 Period: T = 180°
Notice that the amplitude of f(x)=-2sinx is still 2.
32
 THE MODELING FUNCTION f(x) = AsinBx + C
For the function
f(x) = AsinBx + C
we have
The central axis is y=C
The amplitude is |A|
360
The period is T=
B
Notice:
 f(x) ranges between the values C-|A| and C+|A|
 Similar observations apply for f(x) = AcosBx +C
For example, for the function f(x) = 3sin4x+5
The central axis is y=5
The amplitude is 3
360
The period is T=  90
4
and the graph is
33
EXAMPLE 1
The graph of f(x) = Asin(Bx°) + C is given below. Find A, B, C.
The curve is of type +sinx (since y-int central/up)
 Central axis at y=15, so C=15

 Amplitude = 5, so A=5
360 360
 Period T= 8, hence B  = = 45
T 8
Therefore, the equation of the function is f(x) = 5sin(45x°) + 15
EXAMPLE 2
The graph of f(x) = AcosBx +C is given below. Find A, B, C.
90 180
The curve is of type +cosx (since y-int maximum)

ymax  ymin
 Central axis at =5, so C=5
2
 Amplitude = ymax-C =15, so A=15
360
 Period Τ= 180, hence B  2
180
Therefore, the equation of the function is f(x) = 15cos(2x°) + 5
34
Notice that the amplitude is always positive but the coefficient A of

sinx or cosx can be positive or negative.
EXAMPLE 3
Express the following graph as a trigonometric function.
-90 90 180
The curve is of type -sinx (since y-int central/going down)

 Central axis: y=5 hence C=5
 Amplitude = 15 hence A=-15
360
 Period Τ= 180, hence B  2
180
Therefore, the equation of the function is f(x) = -15sin(2x°) + 5
 A POINT SERIES ALMOST SINUSOIDAL

If we are given a series of points with a periodical behavior, for
example,
we can estimate (approximately) a central axis, an amplitude and

a period, so we can find a sinusoidal model function for our data.
NOTICE: The GDC-Regression provides a sinusoidal model of the

form f(x) = a sin(bx+c) + d, but only when x is measured in radians.
So it is not appropriate for the SL syllabus.
35
2.18 VORONOI DIAGRAMS - AN APPLICATION OF LINES
Consider the following simple scenario.
A region contains two sites A and B.
We wish to separate the region into two smaller regions:
 the region of A: all the points which are closest to the site A
 the region of B: all the points which are closest to the site A
It is pretty obvious that the border of the two regions will be the
perpendicular bisector of the line segment AB.
This is perhaps the simplest Voronoi diagram! The two regions are
called cells. The line between the two cells is called an edge.
Well, there is an even simpler case: The trivial Voronoi diagram

with only one site and one cell (which is the whole region).
36
Suppose we have three sites, A, B and C.
If the three sites are collinear (they lie on the same line) the
Voronoi diagram is pretty obvious:
The three edges are parallel.
If, otherwise, the three sites form a triangle, the Voronoi diagram
looks like
Particularly for this case of three sites, we know that the edges
between the three cells are the perpendicular bisectors of the sides.
We know form Geometry that the three edges meet at the same
point. The point P where the edges meet is called a vertex.
We know that all points in cell A all closer to A than to B or C.
37
Let us add one more site D (see below). We obtain four cells and
the edges meet at the two vertices P and Q.
Usually, the sites are given by Cartesian coordinates.
EXAMPLE 1
Suppose that the four sites in the diagram above have coordinates
A(2,3), B(8,3), C(6,7) and D(10,5).
Edge L1 between A and B: it has equation x = 5 (easy!)

Edge L2 between A and C: we find the perpendicular bisector:
6- 2
m AC = = 1 . Hence m  =-1
7- 3
Midpoint AC is M(4,5)
Equation L2: y- 5 =-1(x- 4) , that is y =-x +9
Vertex P: the intersection between L1 and L2. This is P(5,4).
In a similar way, we find:

1 3
Edge L3 between B and C: y = x+
2 2
Edge L4 between B and D: y =-x + 13
23 16
Edge L5 between C and D: y = 2x- 10 , Vertex Q ( , )
3 3
38
Let us have an idea how a more complicated Voronoi diagram looks

like:
Some apparent properties:
 Any point within cell F is closest to F.

 Any point on the edge between cells C and F is equidistant to
sites C and F (and these are the closest sites).
 The vertex between cells C, F and D is equidistant to sites C, F
and D (and these are the closest sites). If we draw a circle with
centre at this vertex, it touches D, F, and D and all other sites
are outside the circle.
 Any vertex connects at least three edges.
39
2.19 PERCENTAGE CHANGE – FINANCIAL APPLICATIONS
In problems with
Interest rates r%
Population growth r%, etc
where
some initial quantity (present value PV)
increases by r% (per year, per month or per any period)
the future value FV after n periods is given by

n
 r 
FV=PV  1  
 100 
If the present value PV decreases by r% per period, the formula

takes the form
n
 r 
FV=PV  1  
 100 
EXAMPLE 1
(a) An amount of 2000 euros is invested at 8% per year. What is
the amount returned after 10 years?
(b) An amount of 2000 euros is depreciated by 8% every year.

What is the amount returned after 10 years?
Solution
For both problems PV=2000, r=8, n=10

n 10
 r   8 
(a) FV=PV  1   =2000  1   =2000(1.08)10  4317.85
 100   100 
n 10
 r   8 
(b) FV=PV  1   =2000  1   =2000(0.92)10  868.78
 100   100 
Remark: In financial mathematics we usually answer in 2dp.
40
Notice: Using GDC (Casio) for Exercise 1

MENU - Financial - F2: [Compound Interest]
n = 10
I% = 8
PV = -2000 (use “-” because we pay)
[Ignore the all the rest; leave PMT=0, P/Y=1, C/Y=1]
Press F5: [FV] and you obtain FV=4317.85
For question (b), set I% =-8. You obtain FV=868.78
r
In fact, this is a geometric sequence with common ratio R=1+
100
Explanation for the common ratio R
Suppose we invest an amount PV for r% per year.
In our example: PV = 2000 r = 8%
In our example In general

Present Value 2000 PV
8 r
Interest 2000× PV×
100 100
Present value + interest
8 r
2000+2000× PV+PV
After 1 year 100 100
 8   r 
2000  1   = PV  1  
 100   100 
r
That is, if an amount increases by r%, we multiply by R=1+
100
2
 r 
After two years we multiply again by R. Thus FV=PV  1  
 100 
n
 r 
Thus after n years: FV=PV  1  
 100 
41
For example, r% is translated into a common ratio as follows:
(increasing case) (decreasing case)

r% r r
R=1+ R=1-
100 100
12% R = 1.12 R = 0.88
20% R = 1.20 R = 0.80
5% R = 1.05 R = 0.95
7.2% R = 1.072 R = 0.928
Be careful of the following slight difference: the initial amount may

be mentioned as value of year 1 ( u1 ) or value of year 0 (PV)
PROBLEM 1 PROBLEM 2 (investment)
Rate of increase 12% Rate of increase 12%

Amount in year 1: u1 =1000. Present value: PV=1000.
Then Then
 In 2nd year u 2 =1000×(1.12)  After 1 year 1000×(1.12)

 In 3rd year u3 =1000×(1.12)2  After 2 years 1000×(1.12)2
 In 5th year u5 =1000×(1.12)4  After 5 years 1000×(1.12)5
 In nth year un =1000×(1.12)n-1  After n years FV=1000×(1.12)n
(Mind that the exponent in Problem 2 is n and not n-1)

In both cases we say that the growth is exponential.
EXAMPLE 2
There are ten boxes in a row. The first box contains 100€ and any
subsequent box contains 10% more that the previous one. What is
the amount in the 10th box?
Solution
Here u1 =100 and r = 1.10. Thus u10 =100(1.10)9  235.8
This is in fact the FV formula, but say that after 9 boxes FV=100(1.10)9
42
EXAMPLE 3 (if the question is about the number of years n)
An amount of 2000 euros is invested at 8% per year. After how

many complete years does the amount exceed 5000?
Solution
n
 8 
FV=2000  1   >5000
 100 
Method A (trial and error): We check several values for n by GDC

and realize that
for n=11 FV = 4663.27

for n=12 FV = 5036.34
Therefore, n = 12
Method B (by using GDC-Financial mode):

Set I% = 8
PV = -2000
FV = 5000 [keep PMT=0, P/Y=1, C/Y=1]
Press F1: [n] and you obtain n=11.9
Since we are looking for complete years we accept the first integer
above 11.9, that is n = 12
Method C (by using SolveN in the GDC):
We can solve the corresponding equation FV = 5000, that is
2000(1.08)n = 5000
The solution is n  11.9. Thus n = 12
Method D (by using logarithms: come back after you learn logs!)
We solve the exponential equation 2000(1.08)n = 5000 by using

logs! The solution is n = log 2.5  11.9.
log 1.08
Thus n = 12
43
EXAMPLE 4
The current population of a city is 800,000. The population
increases by 5.2% every year. Find
(a) the population of the city after 7 years;

(b) the population of the city 7 years ago;
(c) after how many complete years the population of the city
doubles.
Solution
(a) We have an exponential growth with PV=800,000 and r=8%
The population of the city after 7 years is

7
 5.2 
FV=800,000  1    1,140,775
 100 
(b) The formula works for the past as well!
The population of the city 7 years ago was

-7
 5.2 
FV=800,000  1    561,022
 100 
[short explanation:
for the future we multiply by 1.052 every year
for the past we divide by 1.052 every year, or otherwise we
1
multiply by  1.052-1 every year]
1.052
(c) we solve the equation
FV=2×800,000
n
 5.2 
800,000  1   = 1,600,000
 100 
By using a GDC we find n=13.7
Therefore, the population doubles after 14 complete years.
44
 COMPOUNDED INTEREST COMPOUNDED IN k TIME PERIODS
Suppose that an initial amount PV=1000€ is invested with an

interest rate 12% per year. We have seen what happens if the
amount is compounded yearly.
However, the interest may be compounded in k periods per year:
Semiannually (half-yearly): k=2

Quarterly: k=4
Monthly: k=12
For example, if the interest is compounded twice per year

(semiannually), the interest rate for the 6months period is 6% and
Amount after 1 year = 1000×(1.06)2 (1 year = 2 time periods)
Amount after n years = 1000×(1.06)2n
In general, the FV formula takes the form

kn
 r 
FV=PV  1  
 100k 
It is interesting to see how the final amount varies after 5 years,

for the initial amount of 1000 euros: if the interest is compounded
a) Yearly: FV=1000×(1.12)5 = 1762

b) Half-yearly: FV=1000×(1+0.06)2×5 = 1791
c) Quarterly: FV=1000×(1+0.03)4×5 = 1806
d) Monthly: FV=1000×(1+0.01)12×5 = 1817
NOTICE for GDC-Financial mode
The number of periods is denoted by C/Y.
Check the values of FV for

n = 5 PMT=0
I% = 12 P/Y =1
PV = -1000 C/Y = 1, 2, 4, 12 respectively
45
 INVESTMENT WITH REGULAR PAYMENTS
For the rest of the paragraph we mainly use GDC-Financial mode.

However, it is worth to know how the results can be derived.
r
PV is invested with an annual interest rate r% : R  1 
100
An extra PMT is invested at the end of each year
The value of the investment after n years is given by
 Rn  1 
FV  PV  R n  PMT   

 R 1 
Indeed,
n
 r 
- PV is invested for n years: PV  1    PV  R
n
[1st box]
 100 
- 1st PMT is invested for n-1 years: PMT  R n-1

- 2nd PMT is invested for n-2 years: PMT  R n- 2
… …
- last but one PMT is invested for 1 year: PMT  R
- the last payment is PMT PMT
The sum of the last n terms is a G.S. with u 1  PMT , ratio=R

n
R 1 
Thus S n = PMT    [2nd box].

 R  1 
If each payment is equal to the present value (PMT=PV) we obtain

in fact a G.S. of n+1 terms:
 Rn  1   R n 1  1 
FV  PMT  R n  PMT     PMT  
  R 1 

 R 1   
NOTICE
 If we don’t wish to include the last payment in our calculations
we just subtract one payment (PMT)
 In GDC–Financial mode: Insert PMT = negative value [we pay]
46
EXAMPLE 5
An initial amount of 1000 euros and then an extra amount of
1000 euros at the end of each year is invested with an interest
rate 12% compounded yearly (so R =1.12).
The value of the investment after 7 years is
 1.128  1 
FV  1000     12299.69

 1.12  1 
By using GDC–Financial mode
n = 7 PMT= -1000
I% = 12 P/Y =1
PV = -1000 C/Y = 1
FV gives 12299.69.
------------------
Remark: if the last payment is not included we subtract one PMT:
Value = FV-PMT = 12299.69-1000 = 11299.69
EXAMPLE 6
An amount of 1000 euros is invested with an interest rate 12%
compounded yearly (so R =1.12). An extra payment of 300 euros
is added at the end of every year.
 1.127  1 
FV  1000  1.127  300     5237.38

 1.12  1 
n = 7 PMT= -300
I% = 12 P/Y =1
PV = -1000 C/Y = 1
FV gives 5237.38.
------------------
Value = FV-PMT = 5237.38-1000 = 4937.38
47
NOTICE.
If the annual interest rate r% is compounded in k periods then the
ratio for 1 year is
k
 r 
R  1  
 100k 
The formulas for FV above are still valid.
EXAMPLE 7
compounded monthly. An extra payment of 300 euros is added at
the end of each year.
12
 0.12 
R  1   =1.0112=1.12682503
 12 
 R7  1 
FV  1000  R 7  300     5397.73

 R 1 
n = 7 PMT= -300
I% = 12 P/Y =1
PV = -1000 C/Y = 12
FV gives 5397.73.
------------------
Value = FV-PMT = 5397.73-300 = 5097.73
NOTICE.
If the annual interest rate r% is compounded in k periods and the
regular payments also take place in k periods then
 r 
R  1  
 100k 
But now n is the total number of periods, i.e. n=k(years).
48
EXAMPLE 8
compounded monthly. An extra payment of 300 euros is added at
the end of every month.
 0.12 
R  1   =1.01
 12 
The value of the investment after 7 years (so n=712=84) is
 1.0184  1 
FV  1000  1.0184  300     41508.41

 1.01  1 
n = 712 = 84 PMT= -300
I% = 12 P/Y =12
PV = -1000 C/Y = 12
FV gives 41508.41.
------------------
Remark: if the last payment is not included, we subtract one PMT:
Value = FV-PMT = 41508.41-1000 = 41208.41
 ANNUITY – AMORTIZATION
We invest an amount PV but we regularly withdraw an amount
PMT. The only difference is that
we subtract the part of the payments
Now
 Rn  1 
FV  PV  R n  PMT   

 R 1 
Notice for GDC–Financial mode

 We insert the payment PMT as a positive value [we receive]
 Here, the last payment is always included, so we accept FV as is.
49
EXAMPLE 9
compounded monthly. A withdrawal of 150 euros is made at the
end of each year
12
 0.12 
R  1   =1.0112=1.12682503
 12 
 R7  1 
FV  1000  R 7  150    761.22

 R 1 
n = 7 PMT= 150
I% = 12 P/Y =1
PV = -1000 C/Y = 12
FV gives 761.22
If
(annual withdrawal) > (the annual interest)
FV will be zeroed out after a certain time.
Amortization refers to the time needed for this to occur.
In our example above, we solve for n the equation FV=0, i.e.

 Rn  1 
1000  R n  150  

 R 1 
The solution is 15.64 so the last withdrawal will take place after
16 years.
Set FV =0 and press n. We obtain n=15.64, so n=16
For n=15, the GDC gives FV=87.11
For n=16, the GDC gives FV=-51.84
Therefore, the amount left for the last withdrawal is
150-51.84=98.16 euros.
50
ONLY FOR
HL
51
52
2.20 LOGISTIC MODEL

The function of this model (called logistic) has the form
L
f(x) 
1  Ce kx
The three parameters L, C k are all positive.
Consider for example the function

100
y
1  4e  2x
Look at the graph of this function on your GDC with V-window:
5<x<5 and -20<y<70
It looks like exponential. But it is not!
If you increase the upper bound for y you will realise that the
behaviour of graph changes dramatically!
After a while this S-shape curve tends to be constant. In fact it has

a horizontal asymptote at y=100.
53
This function is called logistic and it is the model in many real life
situations.
The spread of a virus in earth is usually exponential in early stages.

Theoretically it tends to + as the time goes by. However, the
population in earth is not infinite, so there is an upper limit for
this spread.
The graph of a logistic model
L
y
1  Ce kx
looks like
The 3 parameters of the function determine the following
it determines the upper bound of the growth

L
horizontal asymptote: y=L
L
C it determines the y-intercept. For x=0, y 
1C
k it determines the growth rate or the steepness of the curve
In our example
100
y
1  4e  2x
 the horizontal asymptote is y=100. Indeed,
x tends to +  e 2x tends to 0  y tends to 100

100
 The y intercept (for x=0) is y   20
14
54
EXAMPLE 1
L
The logistic model P  has a limiting value P=1000.
1  Ce kt
We also know the data
t 0 4
P 160 500
Clearly L=1000.
For t=0, P=160:
L 1000
 160   160  1+C = 6.25  C = 5.25
1C 1C
For t=4, P=500:
1000
 500  1000  1  5.25e  4k  5.25e 4k  1
1  5.25e- 4k 500
ln5.25
 e 4k  5.25  4k  ln5.25  k   0.415
4
Therefore,
1000
P
1  5.25e 0.415t
The value of P exceeds 800 when t = 7.34 (use GDC to confirm).
 A POINT SERIES WHICH IS ALMOST LOGISTIC - REGRESSION
Well, consider the two points of EXAMPLE 1 above and the point
(100,1000) [as a limiting value]. Use GDC: REG: Logistic, to find
the values of L, C and k of the logistic model!
55
2.21 NATURAL LOGARITHMIC MODEL

This model has the form
y  a  blnx
It has two parameters to be determined, a and b.

The graph of such a function has the form
The y-axis (line x=0) is a vertical assympote.

As x increases the value of y also increases but in a slow (and
concave down) way.
So a series of data as follows
seems to follow a natural logarithmic model.

If they give us two points we are able to find the parameters a, b.
56
EXAMPLE 1
Let P  a  blnQ be a model with
Q 4 8
P 15 18
Then
a+bln4=15 (1)
a+bln8=18 (2)
This is in fact a system of two linear equations.
Method A: Directly by GDC-Equation-Simultaneous, with entries
1 ln4 15 
 
1 ln8 18 
we obtain a= 9 and b=4.33. Thus

P  9  4.33lnQ
Method B: Manually (up to a certain point!)

3
(2)-(1) gives: b(ln8-ln4) = 3  bln2 =3  b =
ln2
3ln4
(1) gives: a = 15-  a = 15- 6  a = 9
ln2
Thus
3
P 9 lnQ [which is P  9  4.33lnQ as above]
ln2
 A POINT SERIES ALMOST LOGARITHMIC - REGRESSION
Given a series of points where the trend looks like a natural

logarithmic model, we may use GDC: REG: log to find the model.
For practice, use the two points above to find the model
P  9  4.33lnQ
[Again, for two points only the model is perfect! But given a series
of more points, regression provides the best possible log model]
57
 A COMPARISON OF THREE MODELS

Mind the differences:
An exponential model increases very rapidly.
A power function model increases rapidly but less than exponential.
A natural logarithmic increases slowly.
For example, look at the graphs of the functions
y  ex (exponential)
3
yx (power)
y  20  100lnx (logarithmic)
As you can see the graphs of the exponential and power models
look alike. Thus if we are given some series of points as follows
We can assume that the first model is logarithmic (as it grows

slowly) but for the second one it is not clear at first glance if it is
an exponential or a power function model.
The following paragraph will help.
58
2.22 LINEARISATION OF THREE MODELS
 LINEARISATION OF THE LOGARITHMIC MODEL
Consider the logarithmic model: y  a  blnx
If we set X=lnx, the model takes the form: y  a  bX [linear]
EXAMPLE 1
the function y  3  5lnx
takes the form y  3  5X [where X=lnx]
which is linear!
Let us select some points of y  3  5lnx :
x 2 3 5 10
y 6.47 8.49 11.0 14.5
If X=lnx then
X=lnx 0.693 1.10 1.61 2.30

y 6.47 8.49 11.0 14.5
Observe the corresponding scatter diagrams (i.e. only points)
y in terms of x y in terms of lnx
You see that the points on the second diagram lie on a line.
59
 LINEARISATION OF THE EXPONENTIAL MODEL

Consider the exponential model: y  Ae kx
Apply ln( ): lny = lnA + kx
If we set Y=lny the model takes the form: Y  lnA  bx [linear]
EXAMPLE 2
the function y  3e 2x
takes the form Y  ln3  2X [where Y=lny]
which is linear!
2x
Let us select some points of y  3e :
x 1 2 3 5
y 22.2 163.8 1210.3 66079.4
If Y=lny then
x 1 2 3 5
Y=lny 3.1 5.1 7.1 11.1
Observe the corresponding scatter diagrams:
60
 LINEARISATION OF THE POWER FUNCTION MODEL

Consider the power function model: y  kx n
Apply ln( ): lny = lnk + nlnx
If we set X=lnx and Y=lny it takes the form: Y  lnk  nX [linear]
EXAMPLE 3
the function y  3x 5
takes the form Y  ln3  5X [X=lnx, Y=lny]
which is linear!
5
Let us select some points of y  3x :
x 1 2 3 5
y 3 96 729 9375
If X=lnx, and Y=lny then
X=lnx 0 0.693 1.10 1.61

Y=lny 1.10 4.56 6.59 9.15
Observe the corresponding scatter diagrams:
So now, if we are given a series of points we can investigate by

using linearization which model is the appropriate.
61
EXAMPLE 4
Consider the data
t 3 6 10 12
P 109.417 119.722 134.986 143.333
Investigate by using linearization an appropriate model.

Solution
 We find the corresponding values of lnt and Y=lnP:
t 3 6 10 12
lnt ln3 ln6 ln10 ln12
P 109.417 119.722 135.986 143.333
lnP ln(109.417) ln(119.722) ln(135.986) ln(143.333)
 We check by GDC the graphs of
(P vs lnt) (lnP vs t ) (lnP vs lnt)
 We observe that the diagram (lnP vs t) looks linear.
 Hence we are looking for a relation lnP = at+b
Let us find the values of the parameters a and b by selecting the

first two points:
t 3 6
lnP ln(109.417) ln(119.722)
 We obtain:
3a + b = ln109.417
6a + b = ln119.722
The solution is (by GDC): a = 0.03 b = 4.60516
 Therefore,
lnP = 0.03t+4.60516
P  e 0.03t  4.60516  e 0.03t e 4.60516  100e 0.03t
P  100e 0.03t (exponential model)
62
2.23 PIECEWISE MODEL
A piecewise function has different expressions in different intervals.

For example,
 -x + 6, 0x<2
 2
f(x) =  x , 2x<4
 256  2-x , 4  x  10

which is linear in the 1st interval, quadratic in the 2nd interval and
exponential in the 3rd interval.
Usually, we require such a function to be continuous, that is the

right endpoint of each interval coincides with the left endpoint of
the next interval.
For example, for the function above:
At x=2: -x+6 gives 4 x 2 also gives 4
At x=4: x 2 gives 16 256  2-x also gives 16
Indeed, the curve of the function seems to be “continuous”:
Therefore, given some data with a different behaviour in different

intervals we can constract a piecewise model function to describe
this varying behavior.
63
EXAMPLE 1
Peter investigated a series of data (x,y) and observed that
between x=0 and x=10, y is proportional to x
i.e data follow a model of the form ax
between x=10 and x=30, y decreases linearly in rate 2
i.e data follow a model of the form -2x+b
between x=30 and x=50, data follow the quadratic model
0.1(x- 30)2  10,
(a) Write down a piecewise function f(x) that describes the data.
(b) Find the values of a and b given that f(x) is continuous.
(c) Sketch the graph of the function.
Solution
 ax, if 0  x  10

(a) f(x)   -2x + b, if 10  x  30
 0.1(x- 30)2  10, if 30  x  50

(b) At x=10: ax =-2x + b  10a=-20+b
At x=30: -2x + b = 0.1(x- 30)2  10  60+b=10
The solution of the system isis b=70 and a=5.
 5x, if 0  x  10

(c) f takes the form f(x)   -2x +70, if 10  x  30
 0.1(x- 30)2  10, if 30  x  50

and the graph is
64

Topic 2B. Modelling With Functions

Uploaded by

Copyright:

Available Formats

Topic 2B. Modelling With Functions

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Topic 2B. Modelling With Functions

Uploaded by

Copyright:

Available Formats

International Baccalaureate

2B. Modelling with functions

2.13 LINEAR MODEL ..……………………….…………………………………………………………… 1

2.14 QUADRATIC AND CUBIC MODELS …………………………………………….………… 6

2.15 POWER FUNCTION MODEL – DIRECT/INVERSE VARIATION ……….….. 13

2.16 EXPONENTIAL MODEL ..………………………………………………………………………… 20

2.17 SINUSOIDAL MODEL .………….……………………..….……….……………………….…..…. 30

2.18 AN APPLICATION OF LINES – VORONOI DIAGRAMS ………………..………. 36

2.19 PERCENTAGE CHANGE – FINANCIAL APPLICATIONS ………………………. 40

2.20 LOGISTIC MODEL ……………………………………………………………………………………. 53

2.21 NATURAL LOGARITHMIC MODEL …………………………………………………….…... 56

2.22 LINEARISATION OF THREE MODELS …………………………………………………… 59

2.23 PIECEWISE MODEL ………………………………………………………………………………… 63

2.13 LINEAR MODEL

 CONNECTION WITH ARITHMETIC SEQUENCES

Let us represent the terms of this sequence in the Cartesian plane.

We observe that the points lie on a line. This is not an accident.

The gradient m of the line is equal to the common difference d=3.

 A POINT SERIES ON A LINE

Consider the following data

Let us place the points (x,y) on the Cartesian plane

It seems that the points lie on a line. Indeed,

The gradient of the line segment AB is

The gradient of the line segment BC is

Even for non-consecutive points, for example C and F,

What is the equation of this line?

Let us select the endpoints A(1,6.1) and F(12,18.2).

The gradient of the line segment AF is

Thus, the equation of the line (in point-gradient form) is

We carry on finding the gradient-intercept form:

The equation of the line has the form

We need to find the coefficients a and b:

By substituting A(1,6.1) we obtain a+b=6.1

By substituting F(12,18.2) we obtain 12a+b=18.2

The solution of the system is a=1.1 and b=5.

Therefore, the line is

 A POINT SERIES ALMOST ON A LINE

A first thought to model this linear relation could be to pick two

The solution is a=2, b=-5, hence

If we add the line y=2x-5 on the graph above we obtain

Our model seems to fit quite well to the given points.

 REGRESSION – BEST FIT LINE BY USING GDC

Later, in Statistics (Topic 4), we will discuss in further detain the

MENU: Statistics: Enter List 1: Q-values List 2: P-values

CALC: REG: X [F1]: ax+b [F1]

The best fit line is

2.14 QUADRATIC AND CUBIC MODELS

 POLYNOMIALS AND GRAPHS

Remember the following functions

Linear function y=ax+b

We also say called them polynomials of degree 1 and 2 respectively.

Along the same lines we define the polynomial of degree 3 which

Cubic function y=ax3+bx2+cx+d

The simplest polynomials up to degree 3 are the following

Line quadratic cubic

y=-x y=-x2 y=-x3

For any linear function y=ax+b with a≠0

For any quadratic function y=ax2+bx+c with a≠0

concave down if a<0