Lecture 8

Polynom als

Polynomials
A polynomial (many terms) s a function that has the following
general form of summation of terms.

where
a0 , .... , an are the coefficients,
x is the unknown scalar variable,
n s the degree of polynomial.

2
 Using ezplot command
to plot polynomial functions

The ezplot command (easy plot) draws a symbol c funct on y = f(x),

for x values between -2π and +2π by default.

Degree of Example polynomial Plotting with

polynomial functions Matlab ezplot command

1 y = x+1 ezplot ('x + 1')

2 y = x2 + x + 1 ezplot ('x^2 + x + 1')

3 y = x3 + x 2 + x + 1 ezplot ('x^3 + x^2 + x + 1')

4 y = x4 + x 3 + x 2 + x + 1 ezplot ('x^4 + x^3 + x^2 + x + 1')

Example Plots of 1st, 2nd, 3rd, 4th Degree Polynomials

y=x+1 y = x2 + x + 1

ezplot ('x ^ 3')

y = x3 + x2 + x + 1 y = x4 + x3 + x2 + x + 1

4
 Defining Polynomials as Vectors in Matlab

 In Matlab, polynom als can be def ned as normal vectors.

 All of the coefficients ( nclud ng zeros) of polynom al terms are wr tten n vector.
 The terms should be wr tten from left-to-r ght n decreas ng order of degrees.
 Length of the vector s also the degree of polynom al.

Example polynomial Matlab vector representations

functions (Conta ns the coefficients only)

y = 8x + 3 p = [8 3]
y = x2 –
4x + 10 p = [1 -4 10]
y = 6x2 - 150 p = [6 0 -150]
y = x5 + 6x2 - 7x p = [1 0 0 6 -7 0]

Built-in Polynomial Commands

Matlab Command Description

polyval (p, x) Polynomial evaluation for a g ven value

roots (p) Find roots of a polynomial

poly (r) G ve the polynomial with specified roots

conv (p1, p2) Multiply polynomials

deconv (pnum, pdenom) Divide polynomials

polyder (p) Polynomial derivative

polyfit (x, y, ndegree) Polynomial curve fitting

6
 Calculating Value of a Polynomial
for an x value

 The value of a polynomial at a point can be calculated

with the polyval command that has the form

polyval(p, x)

 P is a vector with the coefficients of the polynomial

 X can be a scalar number,
or a variable that has an assigned value,
or a computable expression,
can also be a vector

Example : Us ng the polyval command

for a scalar value
For the polynomial function f(x) = x3 + 3x2 - 5x - 17

We want to calculate f (5).

>> p = [1 3 -5 -17];

>> polyval(p, 5)

ans =

158

8
 Example : Us ng the polyval command
for a vector
 x s a vector, polyval command returns a correspond ng y vector.
 Program plots the x and y vectors.

f(x) = x3 + 3x2 - 5x - 17
p = [1 3 -5 -17];
x = -5 : 0.1 : 5;
y = polyval(p, x);

% Plotting x and y vectors:

plot(x,y)

grid
xlabel('x')
ylabel('y')

Alternat ve plott ng method:

ezplot ('x^3 + 3*x^2 - 5*x - 17')
9

Finding the roots of a polynomial

 The roots command n Matlab can be used to f nd the roots of a polynomial.

 They are the x values for which the polynomial y function is equal to zero.
r = roots(p)

Example: f(x) = x3+3x2-5x-17 Example: f(x) = 4x2+10x-8

>> p = [1 3 -5 -17]; >> roots( [4 10 -8] )

>> r = roots( p ) ans =

-3.1375
r = 0.6375
2.3186
-2.6593 + 0.5099i Two roots found.
-2.6593 - 0.5099i
Three roots found.
Two of them are complex numbers.
10
Determining coefficients of a polynomial
When the roots of a polynomial are known, the poly command
can be used for determining the coefficients of the polynomial.
p = poly(r)

p is a vector conta n ng the coefficients of the polynomial.

r is a vector conta n ng the roots of the polynomial.

Example:
>> r = [-2 2];

>> p = poly(r)

p =
1 0 -4

Correspond ng polynom al s :
f(x) = x2 - 4
11

Add t on and Subtract on of Polynom als

 Two polynomials can be added and substracted
by adding the vectors of the coefficients.
 If orders are not same, we should add zeros (padd ng)
for the m ss ng terms.
 Example:

>> p1 = [3 15 0 -10 -3 15 -40];

>> p2 = [3 0 -2 -6];

>> p1 + p2
??? Error using ==> plus
Matrix dimensions must agree.

>> p1 + [0 0 0 p2]
ans =
3 15 0 -7 -3 13 -46
12
 Multiplication of Polynom als
 Two polynomials can be multiplied with the conv (convolut on) command
which has the follow ng form:
c = conv(a, b)

 c is a vector of the coefficients of the polynomial that is the product

of the multiplication.
 a and b are the vectors of the polynomials that are being multiplied.
 The two polynomials do not have to be of the same order.

13

Multiplication Example

p1 p2
polynom al polynom al

>> p1 = [2 3 1];
>> p2 = [5 -2];

>> p3 = conv(p1, p2)

p3 =
10 11 -1 -2
x3 x2 x1 x0
14
 Division of Polynom als

Suppose all of the follow ngs are polynomials.

N numerator
D denominator (d v sor) = +
Q quotient (result)
R remainder

A polynomial can be d v ded by another with

the deconv (deconvolution) command.

[Q R] = deconv(N, D)

15

Division Example

D v de 2x3+9x2+7x-6 by x+3

>> u = [2 9 7 -6];
>> v = [1 3];

>> [a b] = deconv(u , v)

a =
2 3 -2 Result of d v s on

b =
0 0 0 0 Rema nder

16
 Finding Derivatives of Polynomials
The polyder command s used to f nd the der vat ve of a polynomial.

Or g nal polynom al: >> f = [3 -2 4];

f(x)= 3x2-2x+4
>> k = polyder(f)
Der vat ve of or g nal:
f’(x)= 6x-2 k =
6 -2

17

Curve Fitting
 When N points (xi, yi) are given , we can f nd a polynomial funct on of
degree N-1 that passes through the points.
 The procedure to find a polynomial vector is called Curve Fitting.

Number of
Polynom al funct on Curve F tt ng Type
known points

F rst degree polynom al f tt ng

2 y = ax + b
(L near L ne f tt ng)

3 y = ax2 + bx + c Second degree polynom al f tt ng

4 y = ax3 + bx2 + cx + d Th rd degree polynom al f tt ng

18
 The polyfit command
 The polyfit command (Polynom al F t) finds the coefficients of a
polynomial that approx mately represents the data (x and y vectors).
p = polyfit(x, y, n)

 p is the vector of the coefficients of the polynomial that fits the data

 x is a vector with the horizontal coordinate

 y is a vector with the vertical coordinate

 n is the degree of the polynomial

If n = 1 then it is called as Linear Regression
If n > 1 then it is called as Polynomial Regression

19

The polyval command

 The polyval command (Polynom al Values) uses the coefficients
to find new values of y, that correspond to the known values of x.

newy = polyval (coefficients, x);

 Before constructing a polynomial with the polyval command,

it is a good idea to plot the g ven x and y data.

 Plott ng w ll show the d str but on of data,

so we can est mate the degree of polynom al.

20
 Example Problem

 Suppose the follow ng (x, y) data po nts are g ven.

(0.9 , 0.9)
(1.5 , 1.5)
(3 , 2.5)
(4 , 5.1)
(6 , 4.5)
(8 , 4.9)
(9.5 , 6.3)

 Wr te a Matlab program to f nd a polynom al (th rd degree)

that s fitted to these po nts by using the polyfit command.
 Also plot the original points as scattered drawing, and
the fitted polynomial as normal drawing.

21

Matlab Program
X = [0.9 1.5 3.0 4.0 6.0 8.0 9.5];
y = [0.9 1.5 2.5 5.1 4.5 4.9 6.3];
katsayilar = polyfit(x, y, 3)

xp = 0.9 : 0.1 : 9.5;

% xp is new vector

yp = polyval(katsayilar, xp);
% yp is new vector

plot(x,y,'o', xp,yp)
grid
xlabel('x, xp')
ylabel('y, yp')
legend('Given data', 'Fitted data')

katsayilar =
Screen
output 0.0220 -0.4005 2.6138 -1.4158
x3 x2 x1 x0 22
 Or g nal Data (Scattered) and F tted Curve
(Th rd degree polyf t)

y = 0.0220x3 - 0.4005x2 + 2.6138x - 1.4158


The green curve represents the th rd
degree polynom al funct on found.

The blue dots represent the or g nal
data values (x and y).

23

Finding a Linear Function

(y = a.x + b)

• We can find the linear function that fits the data by using:
p = polyfit(x, y, 1)

• The first element p(1) will be a (slope of l near l ne).

• The second element p(2) will be b.

24
 Example Problem
 Suppose that to calibrate an instrument, measurements have been done.
 The following table shows the measurement readings against the standard values.
 Wr te a Matlab program to apply a f rst-degree polynom al f t.
 Also use the results to plot a linear line.

Standard Measured
values values
0 0.5030
1 0.7229
2 0.7802
5 1.2106
10 1.7607
15 2.4649

25

Matlab Program
• First-degree polynom al f t means f nd ng a l near function: y = ax+b
• We w ll calculate the a and b coeff c ents.

t = [0 1 2 5 10 15];
m = [0.5030 0.7229 0.7802 1.2106 1.7607 2.4649];

% F rst-degree polyf t
p = polyfit(t, m, 1)

x = 0 : 0.1 : 15 ;
y = polyval(p, x) ;

plot(t,m,'o', x,y)

title('Calibration curve')
xlabel('True values')
ylabel('Measured values')

26
 Screen Outputs
Screen output

p =

0.1267 0.5435

Slope (a) of l near l ne

y = ax + b
y = 0.1267x + 0.5435

The green l ne represents the
l near funct on found.

The blue dots represent the
or g nal data values (x and y).
27