See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/322797649

Newton’s Backward Interpolation: Representation of Numerical Data by a

Polynomial Curve

Article · September 2016

CITATIONS READS

19 29,455

2 authors:

Dhritikesh Chakrabarty Biswajit Das

Handique Girls' College 65 PUBLICATIONS 135 CITATIONS
1,090 PUBLICATIONS 6,306 CITATIONS
SEE PROFILE
SEE PROFILE

All content following this page was uploaded by Dhritikesh Chakrabarty on 15 February 2018.

The user has requested enhancement of the downloaded file.

 Intern at ional Journal of Applied Research 20 16; 2(10): 513-517

ISSN Print: 2394-7500

ISSN Online: 2394-5869
Impact Factor: 5.2 Newton’s backward interpolation: Representation of
IJAR 2016; 2(10): 513-517
www.allresearchjournal.com
numerical data by a polynomial curve
Received: 14-08-2016
Accepted: 15-09-2016
Biswajit Das and Dhritikesh Chakrabarty
Biswajit Das
Research Scholar, Assam
Down Town University, Abstract
Panikhaiti, Guwahati, Assam, In order to reduce the numerical computations associated to the repeated application of the existing
India interpolation formula in computing a large number of interpolated values, a formula has been derived
from Newton’s backward interpolation formula for representing the numerical data on a pair of
Dhritikesh Chakrabarty variables by a polynomial curve. Application of the formula to numerical data has been shown in the
Department of Statistics, case of representing the data on the total population of India corresponding as a function of time. The
Handique Girls’ College, formula is suitable in the situation where the values of the argument (i.e. independent variable) are at
Guwahati, Assam, India equal interval.
& Research Guide, Assam
Down Town University, Keywords: Interpolation, newton’s backward interpolation formula, polynomial curve, representation
Panikhaiti, Guwahati, Assam,
of numerical data
India

1. Introduction
Interpolation, which is the process of computing intermediate values of a function from the
set of given values of the function {Hummel (1947), Erdos & Turan (1938) et al}, plays
significant role in numerical research almost in all branches of science, humanities,
commerce and in technical branches. A number of interpolation formulas namely Newton’s
Forward Interpolation formula, Newton’s Backward Interpolation formula, Lagrange’s
Interpolation formula, Newton’s Divided Difference Interpolation formula, Newton’s Central
Difference Interpolation formula, Stirlings formula, Bessel's formula and some others are
available in the literature of numerical analysis {Bathe & Wilson (1976) [1], Jan (1930),
Hummel (1947) et al}.
In case of the interpolation by the existing formulae, the value of the dependent variable
corresponding to each value of the independent variable is to be computed afresh from the
used formula putting the value of the independent variable in it. That is if it is wanted to
interpolate the values of the dependent variable corresponding to a number of values of the
independent variable by a suitable existing interpolation formula, it is required to apply the
formula for each value separately and thus the numerical computation of the value of the
dependent variable based on the given data are to be performed in each of the cases. In order
to get rid of these repeated numerical computations from the given data, one can think of an
approach which consists of the representation of the given numerical data by a suitable
polynomial and then to compute the value of the dependent variable from the polynomial
corresponding to any given value of the independent variable. However, a method/formula is
necessary for representing a given set of numerical data on a pair of variables by a suitable
polynomial. One such formula has been developed in this study. The formula has been
derived from Newton’s backward interpolation formula. The formula obtained has been
applied to represent the numerical data, on the total population of India since 1971, by a
polynomial curve.

2. Newton’s Backward Interpolation Formula

Correspondence Then Newton’s backward interpolation formula is
Biswajit Das
Research Scholar, Assam f (x) = + v +  +  +
 
Down Town University, ………….
Panikhaiti, Guwahati, Assam,  + …………………….+ 
 
India
~ 513 ~
International Journal of Applied Research

where v =

 
= ( + +

 
+
 

+……………… + ……….

(2.1)
This formula can be expressed as
f (x) = Cn + Cn  1 ( + Cn2 + Cn3 +
Cn 4 + ……………………………….. +
C0 ……….
 (2.2)
where Cn =

Cn1 =

Cn2 =


Cn3 =


Cn4 =

…………………....
……………………..

C0 =


3. Representation of Numerical Data by Polynomial Curve:

By algebraic expansion, one can obtain that
( =(
Also,
= x2  .  . +
= x2  ( + ) +
= x2 – ∑ x x+

Again,

= (x2  .  . +

= x3  x2.  x2 . +  x2. + . + . 

= x3  ( + + x2 + ( + + ) 

= x3 – ∑ x x2 + ∑ ∑ xx x–
Similarly,

= {x3  ( + + x2 + ( + + )  }

= x4  ( + + x3 + ( + + x2   x3

+( + + x2 ( + + ) +

= x4  ( + + + x3 + ( + + + + + ) x2 
+ + + ) +

=x
4
– ∑ x x3 + ∑ ∑ xx x2 – ∑ ∑ ∑ xxx x+

…………………………………………………………………………………………………..
…………………………………………………………………………………………………..
~ 514 ~
International Journal of Applied Research

In general,
……….
= xn – ∑ x xn –1 + ∑ ∑ xx xn –2 –
∑ ∑ ∑ xxx x + ………… + (–1)n ( …………. )
=x – ∑
n
x x n –1
+ ∑ ∑ xx x n –2
– ∑ ∑ ∑ xxx x + …………………
+ (–1)n ( …………. )

Now, equation (2.2), can be expressed as

f(x) = Cn + Cn1 + Cn 2 [ 2 – ∑ x x+ ] + Cn3 [ x3 – ∑ x 2
+
∑ ∑ xx x– ] + Cn 4 [x – ∑
4
x 3
+
∑ ∑ xx 2
– ∑ ∑ ∑ xxx x+ ]
……………………………………………………………………………………..

C0 [ xn – ∑ x xn –1 + ∑ ∑ xx xn –2 – ∑ ∑ ∑ xxx x+
+ ………………. + (–1)n ( …………. )]  (3.1)

This is of the form

f(x) = A0 + A1x + A2 x2 + A3 x3 + …………………….. + An xn  (3.2)
Where
A0 =Cn Cn1 + Cn2 – C n3 + Cn 4
……………………… + (–1)n C0 ( …………. )
A1 = Cn1 – Cn2 ∑ x + C n3 ( ∑ ∑ xx)–
Cn 4 (∑ ∑ ∑ x x x ) + ........... – C0 ∑ ∑ ∑ x x x
A2 = Cn2 – Cn3 ( ∑ x + C n 4 ∑ ∑ x x  …………….
+ C0 (∑ ∑ ∑ ∑ )
A3 = Cn3 – Cn 4 ∑ x + ……………………………………...+
C0 (∑ ∑ ∑ ∑ ∑ )
……………………………………………………………………………………
…………………………………………………………………………………….
An = Cn
Equation (3.2), with the coefficients
A0, A1, A2, A3, …………………….., An,
,as defined above, is the required formula for representing a given set of numerical data on a pair of variables by a suitable
polynomial we have aimed at.

Note: The formula is valid for representing a given set of numerical data on a pair of variables by a suitable polynomial under
the following two conditions:
(i) values of the argument are at equal interval
(ii) the value of x corresponding to which the value of y is to be interpolated is in the last half of the series

4. An Example of the Formula:

The following table shows the data on total population of India corresponding to the years:

Year 1971 1981 1991 2001 2011

Total Population 548159652 683329097 846302688 1027015247 1210193422

Taking 1971 as origin and changing scale by 1/10, one can obtain the following table for independent variable x (representing
time) and f(x) (representing total population of India):

Year 1971 1981 1991 2001 2011

xi 0 1 2 3 4
f(xi) 548159652 683329097 846302688 1027015247 1210193422

Here = 0, = 1, = 2, = 3, = 4
f( ) = 548159652, f( ) = 683329097, f( ) = 846302688, f( = 1027015247
f( ) = 1210193422
Difference Table

~ 515 ~
International Journal of Applied Research

x f(x) f(x) 2 f(x) 3 f(x) 4 f(x)

0 548159652
135169445
1 683329097 27804146
162973591  10065178
2 846302688 17738968  5208174
180712559 15273352
3 1027015247 2465616
183178175
4 1210193422

Now, C4 = = 1210193422

C3 = = 183178175

C2 = = = = 1232808
 . .
  
C1 = = = =  2545558.66
 . . .
  
C0 = = = =  217007.25
 . . . .

The polynomial is

f(x) = A0 + A1x + A2 x2 + A3 x3 + A4 x4……………………………………………………………………………....………. (i)

Where A0 = C4 – C3 + C2 – C1 + C0

= 1210193422 – 183178175 × 4 + 1232808 × 4 × 3 + 2545558.66 × 2 × 3 × 4

 217007. 25 × 1 × 2 × 3 × 4
= 1210193422  732712700 + 14793696 + 61093407.84  5208174
= 1286080525.84  737920874
= 548159651.84

A1 = C3  C2 ( + ) + C1 ( + + )  C0 ( + + + )
= 183178175  1232808 × (3 + 4)  2545558.66 × (6 + 8 + 12) + 217007.25 (6 + 8 + 12 + 24)
= 183178175  1232808 × 7  2545558.66 × 26 + 217007.25 × 50
= 183178175  8629656  66184525.16 + 10850362.5
= 194028537.5 74814181.16
= 119214356.34
A2 = C2  C1 ( + + ) + C0 ( + + + + + )
= 1232808 + 2545558.66 (2 + 3 + 4)  217007.25 (2 + 3 + 6 + 8 + 4 + 12)
=1232808 + 2545558.66 × 9  217007.25 × 35
= 1232808 + 22910027.94  7595253.75
= 24142835.94  7595253.75
= 16547582.19
A3 = C1 – C0 + + +
=  2545558.66 + 217007.25 × (1 + 2 + 3 + 4)
=  2545558.66 + 217007.25 × 10
=  2545558.66 + 2170072.5
=  375486.16
A4 = C0 =  217007.25

(i)  f(x) = 548159651.84 + 119214356.34 x + 16547582.19 x2  375486.16 x3 217007.25 x4 ……………………..(ii)

Which is the required polynomial
Now, putting x = 0, 1, 2, 3, 4 in (ii), we have
f(0) = 548159651.84
= 548159652
f(1) = 548159651.84 + 119214356.34 × 1 + 16547582.19 × 1 375486.16 × 1 217007.25 × 1
= 548159651.84 + 119214356.34 + 16547582.19  375486.16 217007.25
= 683921590.37 592493.41
= 683329096.96
= 683329097
f(2) = 548159651.84 + 119214356.34 × 2 + 16547582.19 × 4  375486.16 × 8 217007.25 × 16
= 548159651.84 + 238428712.68 + 66190328.76  3003889.28 3472116
~ 516 ~
International Journal of Applied Research

= 852778693.28 6476005.28
= 846302688
f(3) = 548159651.84 + 119214356.34 × 3 + 16547582.19 × 9  375486.16 × 27 217007.25 × 81
= 548159651.84 + 357643069.02 + 148928239.71 10138126.32 17577587.25
= 1054730960.57 – 27715713.57
= 1027015247
f(4) = 548159651.84 + 119214356.34 × 4 + 16547582.19 × 16  375486.16 × 64 –
217007.25 × 256
= 548159651.84 + 476857425.36 + 264761315.04  24031114.24 55553856
= 1289778392.24 79584970.24
= 1210193422

5. Conclusion
The formula described by equation (3.2) can be used to represent a given set of numerical data on a pair of variables, by a
polynomial.
The degree of the polynomial is one less than the number of pairs of observations.
The polynomial that represents the given set of numerical data can be used for interpolation at any position of the independent
variable lying within its two extreme values.
The approach of interpolation, described here, can be suitably applied in inverse interpolation also.
Newton’s backward interpolation formula is valid for estimating the value of the dependent variable under the following two
conditions:
(i) Values of the argument are at equal interval
(ii) The value of x corresponding to which the value of y is to be interpolated is in the last half of the series
Therefore, the formula derived here is valid for representing a set of numerical data on a pair of variables by a polynomial
under these two conditions only. Consequently, there is necessity of searching for some formula for representing a set of
numerical data on a pair of variables by a polynomial if the value of the independent variable corresponding to which the value
of the dependent variable is to be estimated lies in the last half of the series of the given values, which are at equal interval, of
the independent variable. Moreover, there is also necessity of searching for some formula for representing a set of numerical
data on a pair of variables by a polynomial if the given values of the independent variable are not at equal interval.

Reference
1. Bathe KJ, Wilson EL. Numerical Methods in Finite Element Analysis, Prentice-Hall, Englewood Cliffs, NJ, 1976.
2. Gerald CF, Wheatley PO. Applied Numerical Analysis, fifth ed., Addison-Wesley Pub. Co., MA, 1994.
3. David R Kincard, Ward Chaney E. Numerical analysis, Brooks /Cole, Pacific Grove, CA, 1991.
4. Endre S, David Mayers. An Introduction to Numerical Analysis, Cambridge, UK, 2003.
5. John H Mathews, Kurtis D Fink. Numerical methods using MATLAB, 4ed, Pearson Education, USA, 2004.
6. James B. Scarborough Numerical Mathematical Analysis, 6 Ed, The John Hopkins Press, USA, 1996.
7. Kendall E. Atkinson An Introduction to Numerical Analysis, 2 Ed., New York. 1989.
8. Chapra SC, Canale RP. Numerical Methods for Engineers, third ed., McGraw-Hill, New York, 2002.
9. Conte SD, Carl de Boor. Elementary Numerical Analysis, 3 Ed, McGraw-Hill, New York, USA, 1980.
10. Robert J Schilling, Sandra L Harries. Applied Numerical Methods for Engineers, Brooks /Cole, Pacific Grove, CA, 2000.

~ 517 ~

View publication stats