KEMBAR78
Lecture2 Interpolation | PDF | Finite Difference | Mathematics Of Computing
Scribd
Upload
81 views3 pages

Lecture2 Interpolation

The document discusses the Newton-Gregory forward difference formula for estimating values of an unknown function f(x) at points between given data points. [1] The formula expresses f(xs) as a series involving forward differences of f(x0) weighted by binomial coefficients of s, where xs = x0 + sh and s is between 0 and 1. [2] An example uses the formula to estimate insurance premiums at age 46 based on premiums from ages 45 to 65. [3] The document notes that the formula is most accurate when s is close to 0, so the backward difference formula is preferable when s is close to 1.

Uploaded by

Ayush Jaiswal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
81 views3 pages

Lecture2 Interpolation

The document discusses the Newton-Gregory forward difference formula for estimating values of an unknown function f(x) at points between given data points. [1] The formula expresses f(xs) as a series involving forward differences of f(x0) weighted by binomial coefficients of s, where xs = x0 + sh and s is between 0 and 1. [2] An example uses the formula to estimate insurance premiums at age 46 based on premiums from ages 45 to 65. [3] The document notes that the formula is most accurate when s is close to 0, so the backward difference formula is preferable when s is close to 1.

Uploaded by

Ayush Jaiswal
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 3

Lecture -2

i. Newton-Gregory forward difference formula:


Let ( , ), = 0, 1, 2, … , be the set of (n + 1) values of an unknown function ( ) at
equidistant points xi , i  1, 2, ... , n in the interval [a, b] where h is the spacing interval

i.e. b  a  nh , a  x0 and b  xn . Then

f  xs   f s  E s f0  1    f0
s

Expanding binomially, one gets.


f s  1  sc1   sc2  2  ...  scn  n  ......  f0 ,  | s | 1

(Note: Binomial expansion of 1  x  is convergent for | n |  1 )


n

f s  f0  sc1  f0  sc2  2 f0  ...  scn  n f0  .......... , | s |  1. …(1)


This is known as Newton’s forward difference formula ( N. Fw ) for equal intervals or
Newton-Gregory forward difference formula.

Error term of the formula : sc


n 1 h n1 f 
n 1
  , x0    xn
Example: From the following table of yearly premiums for policies maturing at
quinquennial ages, estimate the premiums for policies maturing at the age of 46 years.

Age x : 45 50 55 60 65

Premium f  x  : 2.871 2.404 2.083 1.862 1.712


Solution: The difference table of the given data is as follows:

46  45
For xs  46, x0  45, h  5 , so xs  x0  sh  s  0.2
5

Substituting s = 0.2 and values of various order differences lying along the forwardly sloped line
as shown in table in N.Fw formula (1), we get

0.2  0.2  1 2 0.2  0.2  1 0.2  2  3


f  46   f  45   0.2 f  45    f  45    f  45 
2! 3!
0.2  0.8  1.8  2.8 4
  f  45 
4!

0.2  0.8
 2.8710  0.2   0.467     0.146 
2!

0.2  0.8  1.8 0.2  0.8  1.8  2.8


   0.046     0.017 
3! 4!

= 2.8710 – 0.0934 – 0.01168 – 0.002208 – 0.0005712 = 2.763 (Approx.)

Note: It can be observed that this formula is appropriate when the value to be interpolate lie
almost in the beginning of the data table. More explicitly, when xs lies almost in the end of the
data table then only first two or three terms of (1) can be taken into consideration to keep | s | < 1
and hence the result will not be sufficiently accurate. To overcome this problem, we go for next
formula named Newton’s Backward difference formula.

You might also like