MTH2212 – Computational Methods and

Statistics

Curve Fitting and Interpolation

Lecture 6:
Interpolation by Spline Functions
 Objectives

 Introduction
 Linear Splines
 Quadratic Splines
 Cubic Splines

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 2

 Introduction

 Higher order polynomials can lead to erroneous results

because of round off error and oscillations  to be
avoided.

 Lower-order polynomial fits should be used.

 Alternative approach is to apply lower-order polynomials

to subsets of data points : spline functions.

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 3

 Introduction

 The drafting technique of using a spline to draw smooth

curves through a series of points. Notice how, at the end
points, the spline straightens out  This is “natural” spline.

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 4

 Introduction

 Notation used to derive splines. Notice that there are n–1

intervals and n data points.

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 5

 Why Splines ?

16th Order
Polynomial

Original
Function

4th Order
Polynomial

8th Order
Polynomial

 Higher order polynomial interpolation is a bad idea

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 6

 Why Splines ?

 f(x) = 1/(1+25x2)

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 7

 Spline Interpolation Definition

 Notation used to derive splines. Notice that there are n–1

intervals and n data points.
 Each interval i has its own spline function Si(x)

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 8

 Spline Interpolation Definition

 Given n distinct knots xi such that:

x1  x2  ...  xn 1  xn

with n knot values fi find a spline function

S1 ( x) x  [ x1 , x2 ]
S ( x) x  [ x , x ]

S ( x)   2 2 3


S n 1 ( x) x  [ x n 1 , x n ]

with each Si(x) a polynomial of degree at most n.

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 9

 Linear Splines

 Simplest form of spline interpolation

 Given the data ( x1 , f1 ), ( x2 , f 2 ),....., ( xn 1 , f n 1 ), ( xn , f n )
 Consecutive data points connected through straight lines.
 Each Si is a linear interpolation function constructed as:

S i ( x)  ai  bi ( x  xi )

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 10

 Linear Splines

 Using the formulation of Newton linear interpolation, the

linear splines can be written as:

 f ( x 2 )  f ( x1 )
S1 ( x)  f ( x1 )  x2  x1
( x  x1 ) x  [ x1 , x2 ]

 f ( x3 )  f ( x 2 )
S 2 ( x)  f ( x 2 )  ( x  x 2 ) x  [ x 2 , x3 ]
S ( x)   x3  x 2


 f ( xn )  f ( xn 1 )
S n 1 ( x)  f ( x n 1 )  xn  xn 1
( x  x0 ) x  [ xn 1 , xn ]


Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 11

 Example 1

 Given the following data, evaluate the function at x=5 using

linear splines.

i xi fi
1 3.0 2.5
2 4.5 1.0
3 7.0 2.5
4 9.0 0.5

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 12

 Example 1 - Solution

 x=5 is between x=4.5 and x=7. Hence, f(5) is evaluated using

linear splines on the second interval
f ( x3 )  f ( x2 )
S 2 ( x )  f ( x2 )  ( x  x2 )
x3  x2
2.5 - 1.0
 1.0  ( x  4.5)
7.0 - 4.5
 The function value at x=5

2.5-1.0
f (5)  S 2 (5)  1.0  (5  4.5)  1.3
7.0-4.5

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 13

 Quadratic Splines

 Quadratic splines are rarely used for interpolation for

practical purposes

 Ideally quadratic splines are only used to understand cubic

splines

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 14

 Quadratic Splines

 Each Si is a second-order polynomial constructed as:

S i ( x)  ai  bi ( x  xi )  ci ( x  xi ) 2

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 15

 Quadratic Splines

 The splines are given by:

S1 ( x)  a1  b1 ( x  x1 )  c1 ( x  x1 ) 2 x  [ x1 , x2 ]

S 2 ( x)  a2  b2 ( x  x2 )  c2 ( x  x2 ) x  [ x2 , x3 ]
2
S ( x)  

S ( x)  a  b ( x  x )  c ( x  x ) 2 x  [ xn1 , xn ]
 n1 n 1 n 1 n 1 n 1 n 1

 Find ai, bi, ci for i=1,2,…,n-1  3(n-1) unknown constants to

evaluate  3(n-1) equations are required.

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 16

 Derivation of Quadratic Splines

1. The function must pass through all the points (Continuity

condition), i.e. for x=xi, we have:
S (
 i i x )  f i  ai  bi ( xi  xi )  ci ( xi  xi ) 2
ai  f i
(n-1) equations
2. Functions value of adjacent polynomials must be equal at
the knots i.e. for knot i+1 we have:
fi  bi ( xi 1  xi )  ci ( xi 1  xi ) 2  fi 1  bi 1 ( xi 1  xi 1 )  ci 1 ( xi 1  xi 1 ) 2
By setting hi = xi+1 – xi, the above equation simplifies to:
2
f i  bi hi  ci hi  f i 1
this equation can be written for knots i=1,2,…,n-1
(n-1) equations
Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 17
 Derivation of Quadratic Splines

3. The first derivatives at the interior nodes must be equal

(adjacent splines will be joined smoothly).
Quadric spline function can be differentiated to
S 'i ( x)  bi  2ci ( x  xi )
first derivatives at interior knot i+1 gives:
bi  2ci hi  bi 1
(n-2) equations
4. Make one arbitrary choice i.e. assume the second derivative
is zero at the first point.
c1  0
(1) equations
Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 18
 Example 2

 Given the following data, evaluate the function at x=5 using

Quadratic splines.

i xi fi
1 3.0 2.5
2 4.5 1.0
3 7.0 2.5
4 9.0 0.5

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 19

 Example 2 - Solution

 Since there are 4 data points, 3 quadratic splines pass

through them
S1 ( x)  a1  b1 ( x  3)  c1 ( x  3) 2 x  [3,4.5]
S 2 ( x)  a2  b2 ( x  4.5)  c2 ( x  4.5) 2 x  [4.5,7]
S3 ( x)  a3  b3 ( x  7)  c3 ( x  7) 2 x  [7,9]
 The necessary function and interval width values are:
i xi fi hi
1 3.0 2.5 4.5 – 3.0 = 1.5
2 4.5 1.0 7.0 – 4.5 = 2.5
3 7 2.5 9.0 – 7.0 = 2.0
4 9 0.5

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 20

 Example 2 - Solution

 Setting up the equations

1. The function must pass through all the points:
a1  f1  2.5
ai  f i  a2  f 2  1.0
a3  f 3  2.5

2. Functions value of adjacent polynomials must be equal at

the knots:
2
f1  b1h1  c1h1  f 2 2.5  1.5b1  1.0
2
f i  bi hi  ci hi  f i 1  f 2  b2 h2  c2 h2 2  f 3  1.0  2.5b2  6.25c2  2.5
f 3  b3h3  c3h3  f 4
2 2.5  2.0b3  4c3  0.5

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 21

 Example 2 - Solution

3. The first derivatives at the interior nodes must be equal

b1  b2 b1  b2
bi  2ci hi  bi 1  
b2  2c2 h2  b3 b2  5c2  b3

4. Assume the second derivative is zero at the first point.

c1  0

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 22

 Example 2 - Solution

 All the equations can be assembled in a matrix form as

1.5 0 0 0 0  b1   1.5
 0 2.5 6.25 0
 0 b2   1.5 
   
0 0 0 2 4 c2     2 
 
 1 1 0 0 0 b3   0  b1  1.0
   
 0 1 5 1 0 c3   0  b2  1.0
 These equations can be solved with the results: c2  0.64
b3  2.2
a1  2.5 c3  1.6

 Along with a2  1.0 and c1  0

a3  2.5
Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 23
 Example 2 - Solution

 The quadratic splines are then given by

S1 ( x)  2.5  ( x  3) x  [3,4.5]
S 2 ( x)  1.0  ( x  4.5)  0.64( x  4.5) 2 x  [4.5,7]
S3 ( x)  2.5  2.2( x  7) - 1.6( x  7) 2 x  [7,9]
 Because x=5 lies in the second interval, we use S2 to make
the prediction
S 2 (5)  1.0  (5  4.5)  0.64(5  4.5) 2  0.66

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 24

 Cubic Splines

 Third-order curves used to connect each pair of data

points are called cubic splines.

si ( x)  ai  bi ( x  xi )  ci ( x  xi ) 2  d i ( x  xi ) 3

 We have n data points  n-1 intervals  4n-4 unknowns

constants  4n-4 conditions are needed

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 25

 Cubic Splines

 The first 3 conditions are identical to those used for the

quadratic splines:
1. Spline functions must pass through all the data points
2. Function values must be equal at the interior knots
3. First derivatives at the interior knots must be equal
4. Second derivatives at the interior knots must be equal
5. Second derivatives at the end knots are zero
Natural Spline
Or
5. First derivatives at the first and last knots are specified
Clamped End

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 26

 Derivation of Cubic Splines

1. The function must pass through all the points (Continuity

condition), i.e. for x=xi, we have:
ai  f i
2. Functions value of adjacent polynomials must be equal at
the knots i.e. for knot i+1 we have:
f i  bi hi  ci hi 2  d i hi 3  f i 1

3. The first derivatives at the interior nodes must be equal i.e.

at interior knot i+1 gives:
2
bi  2ci hi  3d i hi  bi 1

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 27

 Derivation of Cubic Splines

4. The second derivatives at the interior nodes must also be

equal.
Cubic spline function can be differentiated to
S ''i ( x)  2ci  6d i ( x  xi )
second derivatives at interior knot i+1 gives:
ci  3d i hi  ci 1

5. Second derivatives at the end knots are zero

c1  0
cn  0

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 28

 Derivation of Cubic Splines

 Let’s summarize the obtained equations:

ai  f i (1)
f i  bi hi  ci hi 2  d i hi 3  f i 1 (2)
2
bi  2ci hi  3d i hi  bi 1 (3)
ci  3d i hi  ci 1 (4)
c1  0
(5)
cn  0

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 29

 Derivation of Cubic Splines

 Solve Eq. (4) for di:

ci 1  ci
di  (6)
3hi

 Substitute into Eq. (2) and (3)

hi2
f i  bi hi  (2ci  ci 1 )  f i 1 (7)
3
bi 1  bi  hi (ci  ci 1 ) (8)
 Solve Eq.(7) for bi:
f i 1  f i hi
bi   (2ci  ci 1 ) (9)
hi 3

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 30

 Derivation of Cubic Splines

 Reduce the index of Eq. (8) and (9) by 1

f i  f i 1 hi 1
bi 1   (2ci 1  ci ) (10)
hi 1 3

bi  bi 1  hi 1 (ci 1  ci ) (11)

 Substitute Eq. (9) and (10) into Eq. (11)

hi 1ci 1  2(hi 1  hi )ci  hi ci 1  3( f [ xi 1 , xi ]  f [ xi , xi 1 ]) (12)

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 31

 Derivation of Cubic Splines

 Equation (12) can be written in matrix form as

1   c1   0 
h 2(h  h ) h   c   3( f [ x , x ]  f [ x  x ]) 
 1 1 2 2   2   3 2 2 1 
          
 
 hn 2 2(hn 2  hn 1 ) hn 1  cn 1  3( f [ xn , xn 1 ]  f [ xn 1  xn  2 ])
   
 1   cn   0 

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 32

 Example 3

 Given the following data, evaluate the function at x=5 using

Cubic splines.

i xi fi
1 3.0 2.5
2 4.5 1.0
3 7.0 2.5
4 9.0 0.5

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 33

 Example 2 - Solution

 Since there are 4 data points, 3 quadratic splines pass

through them
S1 ( x)  a1  b1 ( x  3)  c1 ( x  3) 2  d1 ( x  3)3 x  [3,4.5]
S 2 ( x)  a2  b2 ( x  4.5)  c2 ( x  4.5) 2  d 2 ( x  4.5)3 x  [4.5,7]
S3 ( x)  a3  b3 ( x  7)  c3 ( x  7) 2  d 3 ( x  7)3 x  [7,9]
 The necessary function and interval width values are:
i xi fi hi
1 3.0 2.5 4.5 – 3.0 = 1.5
2 4.5 1.0 7.0 – 4.5 = 2.5
3 7 2.5 9.0 – 7.0 = 2.0
4 9 0.5

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 34

 Example 2 - Solution

 Determine the c coefficients from the set of simultaneous

equations
1   c1   0 
 h 2( h  h ) h  c  3( f [ x , x ]  f [ x  x ]) 
 1 1 2 2   2    3 2 2 1 

 h2 2( h2  h3 ) h3  c3  3( f [ x4 , x3 ]  f [ x3  x2 ])
    
 1  c4   0

1   c1   0  c1  0
1.5 8 2.5  c   4.8  c2  0.8395
   2    

 2.5 9 2 c3   4.8 c3  0.7665
   
 1  c4   0  c4  0

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 35

 Example 2 - Solution

 Use equations (9) and (6) to compute bi and di

b1  1.42
f i 1  f i hi
bi   (2ci  ci 1 ) b2  0.16
hi 3
d1  0.187 b3  0.022
ci 1  ci
di  d 2  0.214
3hi
d 3  0.128
 The cubic splines for each interval

S1 ( x)  2.5  1.42( x  3)  0.187( x  3)3 x  [3,4.5]

S 2 ( x)  1.0  0.16( x  4.5)  0.84( x  4.5) 2  0.214d 2 ( x  4.5)3 x  [4.5,7]
S3 ( x)  2.5  0.022( x  7) - 0.767( x  7) 2  0.128( x  7)3 x  [7,9]

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 36

 Example 2 - Solution

 The function value at x=5 is calculated using S2

S 2( 5 )  1.0  0.16( 5  4.5 )  0.84( 5  4.5 )2  0.214d 2( 5  4.5 )3
 1.103

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 37

 Example - Summary

 Spline fits of the set of four

points given in the example
(a) Linear spline,
(b) quadratic spline, and
(c) cubic spline, with a cubic

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 38

 Assignment #3

 Computational Methods
 20.10, 20.32
 Statistics
 Check with Dr. Faiz

Dr. M. Hrairi MTH2212 - Computational Methods and Statistics 39