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

net/publication/322797688

Representation of Numerical Data by Some Special Mathematical Curves

Article in International Journal of Electronics and Applied Research · August 2018

DOI: 10.33665/IJEAR.2017.v04i01.002

CITATIONS READS

17 4,993

1 author:

Dhritikesh Chakrabarty
Handique Girls' College
1,164 PUBLICATIONS 6,631 CITATIONS

SEE PROFILE

All content following this page was uploaded by Dhritikesh Chakrabarty on 09 November 2022.

The user has requested enhancement of the downloaded file.

International Journal of Electronics and Applied Research (IJEAR) vol. 4, issue 1, July 2017
Online (http://eses.net.in/ESES Journal) ISSN 2395-0064

Representation of Numerical Data by Some Special Mathematical Curves

Dhritikesh Chakrabarty
Department of Statistics, Handique Girls’ College, Guwahati, Assam, India,
[email protected], [email protected]

Abstract

It is be possible to represent a set of numerical data on a pair of variables by special

mathematical curves (other than polynomial curve) besides the representation of said set of
numerical data by polynomial curve for which several methods have already been
developed. Recently, some studies have been made on mathematical representation of
numerical data on a pair of variables by some special mathematical curves (other than
polynomial curve) namely exponential curve (both simple & modified), Pearl & Reed
curve and Makeham’s curve. Methods have been developed for representing numerical
data by these curves. This paper has been prepared for brief description of the methods for
representing numerical data on a pair of variables by these three special curves with
numerical examples in order to show the applications of the methods to numerical data.

Key words: Pair of variables; special curve; numerical data; method of representation.

1. Introduction

Mathematical representation of numerical data plays significant role in many problems in research and
investigation. There exist statistical methods for representing numerical data by mathematical curve.
However representations by these methods are not accurate but approximate. Recently, some studies have
been made on accurate mathematical representation of numerical data by pol Chakrabarty ynomial curve
[Chakrabarty , 2016a ; Chakrabarty krabarty 2016b ; Chakrabarty 2016c ; Chakrabarty 2016d ;
Chakrabarty 2016e; Chakrabarty 2017a; Chakrabarty 2017b ;Das & Chakrabarty 2016a ; Das &
_____________________________________________________________________________________
Copy right @ESES IJEAR
https://doi.org/10.33665/IJEAR.2017.v04i01.002
52
International Journal of Electronics and Applied Research (IJEAR) vol. 4, issue 1, July 2017
Online (http://eses.net.in/ESES Journal) ISSN 2395-0064

Chakrabarty 2016b ; Das & Chakrabarty 2016c ; Das & Chakrabarty 2016d; Das & Chakrabarty 2016e
;Das & Chakrabarty 2016f ; Das & Chakrabarty 2017a ; Das & Chakrabarty 2017b]. Some formulae have
been derived by them for representation of numerical data by polynomial curve [Das & Chakrabarty,
2016a - 2016f ]. They have also derived the first formula [Das & Chakrabarty, 2016a] from Lagrange’s
Interpolation equation [Echols ,1893; Corliss, 1938 ; Hummel, 1947; Traub, 1964 ; Jardan, 1965;
Quadling, 1966 ; Mills, 1977 ; Kendall, 1989 ; Vertesi, 1990 , Kincard & Ward, 1991 ; Revers & Michael
, 2000 ; Endre & David, 2003 ; Jan, 2004 ; Nasrin et al., 2010]. The second formula [Das & Chakrabartyl,
2016b] has been derived from Newton’s Divided Difference Interpolation Formula [Herbert, 1962 ;
Whittaker & Robinson 1967a , 1967b ; De Boor et al., 2003.]. The third one [Das & Chakrabarty, 2016c]
has been derived from Newton’s Forward Interpolation formula [Erdos & Turan, 1938 ; Whittaker &
Robinson, 1967a ; Nasrin, 2010 et al] and the fourth one [Das & Chakrabarty 2016d] has been derived
from Newton’s Backwards Interpolation formula [Erdos & Turan, 1938 ; Hummel, 1947 ; Jardan ,1965 ;
Whittaker & Robinson, 1967a ; Kendall, 1989 ; Kincard & Ward, 1991 ; Endre & David, 2003 ; Jan et
al., 2004]. Later on one method has been constructed by Chakrabarty 2016a , 2016b , 2016c for
representing a given set of numerical data on a pair of variables by a suitable polynomial. The method
based on two numerical operations namely finite difference operation and ratio operation [Gertrude, 1954
; Herbert, 1962 ; Jordan, 1965 ; Dokken & Lyche 1979 ; Fred, 1979 ; Jeffreys & Jeffreys, 1988 ; Lee,
1989 ; Chwaiger, 1994 ; De Boor, 2003 ; Endre & David, 2003 ; Floater et al., 2003]. In another study,
three methods have been composed by Das & Chakrabarty [2016e ; 2016f ; 2017a] for the same purpose.
The two methods are based on the inversion of a square matrix. The first one is based on matrix inversion
from Cayley-Hamilton theorem [Cayley, 1858 , 1889 ; Hamilton et al., 1853 ; 1862 ; 1864a ; 1864b.]
while the second & the third ones are on matrix inversion by elementary row transformation &
elementary column transformation of matrix [Cayley et al., 1858 , 1889 ; Hamilton et al., 1889]
respectively.
It is be possible to represent the numerical data on a pair of variables by special mathematical curves
other than polynomial curve besides the representation of said numerical data by polynomial curve for
which several methods have already been developed. Recently, some studies have been made on
mathematical representation of numerical data on a pair of variables by some special mathematical curves
other than polynomial curve namely exponential curve (both simple & modified) [Abramowitz & Stegun,
1972; Beyer, 1987 ; Krantz ,1999] , Pearl & Reed curve [Gershenfeld, 1999 ; Jannedy et al., 2003] and
_____________________________________________________________________________________
Copy right @ESES IJEAR

53
International Journal of Electronics and Applied Research (IJEAR) vol. 4, issue 1, July 2017
Online (http://eses.net.in/ESES Journal) ISSN 2395-0064

Makeham curve [Makeham , 1860; Kenney & Keeping, 1962 ; Makeham, 1874]. Methods have been
developed for representing numerical data by these curves. This paper has been prepared for brief
description of the methods for representing numerical data on a pair of variables by these three special
curves with numerical examples in order to show the applications of the methods to numerical data.

2. Representation of Numerical Data by Exponential Curve

The exponential curve is of the form

y=a (2.1)
where a & b are parameters.
Equation (2.1) implies,
log y = log a + x log b (2.2)
Since there are two parameters in the exponential curve, two equations are necessary for determining the
values of the parameters and accordingly two sets of values the pair of variables are necessary.

Let & be the values of y corresponding to the values & of x respectively.

Then the points

lie on the equation (2.1) and hence they satisfy (2.2).

Accordingly,

log = log a + log b (2.3)

& log = log a + log b (2.4)

(2.4)  (2.3) 

= log b ( )

 log b = ) i.e. b = antilog ) (2.5)

Now, (2.3) & (2.4) 

log a = log  ( = log  )

_____________________________________________________________________________________
Copy right @ESES IJEAR

54
International Journal of Electronics and Applied Research (IJEAR) vol. 4, issue 1, July 2017
Online (http://eses.net.in/ESES Journal) ISSN 2395-0064

i.e. a = antilog{log  )}

= antilog{log  )} (2.6)

2.1. Example of Data Representation by Exponential Curve

The following table shows the data on total population of India corresponding to the years:
Table-2.1.1
Year (t) Total Population P(t) Year (t) Total Population P(t)
1951 361088090 1981 683329097
1961 439234771 1991 846302688
1971 548159652 2011 1210193422

Let us first represent the total populations corresponding to the years 1951 & 1961 by the exponential
curve described by equation (2.1).
In order to do it let us take the year 1951 as origin (i.e 0) and choose the scale 1/10 such that the value of
x corresponding to the year 1961 becomes 1. Thus we have the following table:
Table-2.1.2
T x P(t) log P(t)  log P(t)
1951 0 361088090 19.704632503150 0.195912110819
1961 1 439234771 19.900544613969

Now by (2.5), b = 1.21641999047

Accordingly by (2.6), a = 361088090
Therefore, the exponential curve that can represent the total populations corresponding to the years 1951
& 1961

P(t) = 361088090 × = 361088090 ×

This curve yields,
P(1951) = 361088090 × = 361088090

_____________________________________________________________________________________
Copy right @ESES IJEAR

55
International Journal of Electronics and Applied Research (IJEAR) vol. 4, issue 1, July 2017
Online (http://eses.net.in/ESES Journal) ISSN 2395-0064

& P(1961) = 361088090 × = 439234771

Thus, the values of P(1951) & P(1961) yielded by the equation of the exponential curve are identical to
the corresponding given values of them.
Similarly, the data on total population corresponding to two consecutive years (at a gap of 10 years, 20
years, 30 years etc.) can be represented by the exponential curve. The curves obtained have been shown
in the following table (Table-2.1.3).
Table-2.1.3
Exponential curve representing P(t) total population P(t) of India

Years (t) Equation of the exponential curve representing P(t)

1951 , 1961 P(t) = 361088090 ×

1961 , 1971 P(t) = 439234771 ×
1971 , 1981 P(t) = 548159652 ×
1981 , 1991 P(t) = 683329097 ×
1991 , 2001 P(t) = 846302688 ×
2001 , 2011 P(t) = 1027015247×
1951 , 1971 P(t) = 361088090 ×
1961 , 1981 P(t) = 439234771 ×
1971 , 1991 P(t) = 548159652 ×
1981 , 2001 P(t) = 683329097 ×
1991 , 2011 P(t) = 846302688 ×
1951 , 1981 P(t) = 361088090 ×
1961 , 1991 P(t) = 439234771 ×
1971 , 2001 P(t) = 548159652 ×
1981 , 2011 P(t) = 683329097 ×

_____________________________________________________________________________________
Copy right @ESES IJEAR

56
International Journal of Electronics and Applied Research (IJEAR) vol. 4, issue 1, July 2017
Online (http://eses.net.in/ESES Journal) ISSN 2395-0064

1951 , 1991 P(t) = 361088090 ×

1961 , 2001 P(t) = 439234771 ×
1971 , 2011 P(t) = 548159652 ×
1951 , 2001 P(t) = 361088090 ×
1961 , 2011 P(t) = 439234771 ×
1951 , 2011 P(t) = 361088090 ×

3. Representation of Numerical Data by Modified Exponential Curve

The modified exponential curve is of the form

y=a+b (3.1)
where a , b & c are parameters.
Let

,
be the values of y corresponding to the values of

of x respectively.
Then the points

( , ),( , ),( , )
lie on the curve described by equation (2.1).
Therefore,

=a+b , i= 0,1,2 (3.2)

which implies,

=b(  ) , i= 0,1 (3.3)

If , , are equally spaced then

 =  =h

_____________________________________________________________________________________
Copy right @ESES IJEAR

57
International Journal of Electronics and Applied Research (IJEAR) vol. 4, issue 1, July 2017
Online (http://eses.net.in/ESES Journal) ISSN 2395-0064

i.e. = +h & = + 2h
Then,

=a+b , i= 0,1,2 (3.4)

Accordingly,

=b ( & =b ( (3.5)
Which implies,

i.e.

i.e. c = antilog { (3.6)

From equations in (3.5) expression for b is obtained as

b= = (3.7)
where c is given by equation (3.6).
Finally from equations in (3.4), expression for a is obtained as

a= b = b = b (3.8)
where b & c are given by equations (3.7) & (3.6) respectively.

3.1. Example of Data Representation by Modified Exponential Curve

Let us consider the data shown in Table-2.1.1 mentioned in section 2.1.

Let us first represent the total populations corresponding to the years 1951, 1961 & 1971 by the
modified exponential curve described by equation (3.1).
In order to do it let us take, as earlier, the year 1951 as origin (i.e. 0) and choose the scale 1/10 such
that the value of x corresponding to the year 1961 becomes 1 and 1971 becomes 2. Thus we have the
following table (Table-3.1.1).

_____________________________________________________________________________________
Copy right @ESES IJEAR

58
International Journal of Electronics and Applied Research (IJEAR) vol. 4, issue 1, July 2017
Online (http://eses.net.in/ESES Journal) ISSN 2395-0064

Table-3.1.1

t x P(t) P(t)

1951 0 361088090 78146681 1.393851659547 0.332070893151

1961 1 439234771 108924881
1971 2 548159652

Now by (3.6), c = 1.3938516595

Accordingly by (3.7), b = 198416533.52
and by (3.8), a = 162671556.48

Therefore, the modified exponential curve that can represent the given data is

P(t) = 162671556.48 + 198416533.52 ×

This curve yields,

P(1951) = 162671556.48 + 198416533.52 = 361088090,

P(1961) = 162671556.48 + 198416533.52 × 1.3938516595 = 439234771
& P(1971) = 162671556.48 + 198416533.52 × 1.3938516595 2 = 548159652

Similarly, the data on total population corresponding to two consecutive years (at a gap of 30 years)
can be represented by the exponential curve. The curves obtained have been shown in the following tables
Table-3.1.2.
_____________________________________________________________________________________
Copy right @ESES IJEAR

59
International Journal of Electronics and Applied Research (IJEAR) vol. 4, issue 1, July 2017
Online (http://eses.net.in/ESES Journal) ISSN 2395-0064

Table-3.1.2
Modified exponential curve representing P(t) total population P(t) of India

Years (t) Equation of the modified exponential curve representing P(t)

1951 , 1961 , 1971 P(t) = 162671556.48 + 198416533.52 ×

1961 , 1971 , 1981 P(t) =  12844741.85 + 452079512.85 ×
1971 , 1981 , 1991 P(t) =  108964607.87 + 657124259.87 ×
1981 , 1991 , 2001 P(t) =  813961579.52 + 1497290676.52 ×
1991 , 2001 , 2011 P(t) =  12398675095.11 + 13244977783.11

×
1951 , ,1981 , 2011 P(t) =  146377372.04 + 507465462.04 ×

4. Representation of Numerical Data by Pearl & Reed Curve

The Pearl & Reed curve is of the form

y= (4.1)
where , B and C are parameters
As earlier, let

,
be the values of y corresponding to the values of

, ,
of x respectively.
Then the points
( , ),( , ),( , )
lie on the curve described by equation (4.1).
_____________________________________________________________________________________
Copy right @ESES IJEAR

60
International Journal of Electronics and Applied Research (IJEAR) vol. 4, issue 1, July 2017
Online (http://eses.net.in/ESES Journal) ISSN 2395-0064

Therefore,

= i.e. = + , i=0,1,2 (4.2)

This  =  , i=0,1 (4.3)

If , , are equally spaced then

 =  =h i.e. = +h & = + 2h
Thus,

= (  1) , i=0,1 (4.4)
which implies,

log { } = h log C (4.5)

so that

log C = log { }

i.e. C = antilog [ log { } (4.6)

From equations (4.4), expression for A is obtained as

A= = (4.7)
where C is given by equation (4.6).
Finally from equations (4.3), expression for B is obtained as

B=  =  =  (4.8)
where A & C are given by equations (4.7) & (4.6) respectively.

4.1. Example of Data Representation by Pearl & Reed Curve

_____________________________________________________________________________________
Copy right @ESES IJEAR

61
International Journal of Electronics and Applied Research (IJEAR) vol. 4, issue 1, July 2017
Online (http://eses.net.in/ESES Journal) ISSN 2395-0064

Let us first represent the total populations corresponding to the years 1951, 1961 & 1971 by the Pearl &
Reed curve described by equation (4.1).
In order to do it let us take, as earlier, the year 1951 as origin (i.e. 0) and choose the scale 1/10 such
that the value of x corresponding to the year 1961 becomes 1 and 1971 becomes 2. Thus we have the
following table (Table-4.1.1).
Table-4.1.1

t X P(t) =
 ( )

1951 0 361088090 0.000000002769407321


1961 1 439234771 0.000000002276686788 0.000000000492720533 0.9181690619

1971 2 548159652 0.0000000018242860384 

0.0000000004524007496

Now,
C = 0.9181690619
Accordingly,

A = = = 166079821.3578811825

& B=  =  1 =  0.5400573269

Therefore, the Pearl & Reedcurve that can represent the given data is

y=

_____________________________________________________________________________________
Copy right @ESES IJEAR

62
International Journal of Electronics and Applied Research (IJEAR) vol. 4, issue 1, July 2017
Online (http://eses.net.in/ESES Journal) ISSN 2395-0064

i.e. P(t) =
This curve yields,

P(1951) = = =
= 361088090 ,

P(1961) = = =
= 439234771

& P(1971) = = =
= 548159652
These values are identical with the corresponding observed values shown in Table-2.1.1 mentioned in
section 2.
The data on total population corresponding to three consecutive years (at a gap of 10 years) can be
represented by the Pearl & Reed curve. The curves obtained have been shown in the following table
(Table-4.1.2).
Table-4.1.2
Pearl & Reed curve representing total population P(t) of India

Years (t) Equation of the Pearl & Reed curve representing P(t)
1951
1961
1971 P(t) =

1961
1971
1981 P(t) =

1971

_____________________________________________________________________________________
Copy right @ESES IJEAR

63
International Journal of Electronics and Applied Research (IJEAR) vol. 4, issue 1, July 2017
Online (http://eses.net.in/ESES Journal) ISSN 2395-0064

1981
1991
P(t) =
1981
1991
2001 P(t) =

1991
2001
2011 P(t) =

1951
1981
2011
P(t) =

5. Representation of Numerical Data by Makeham’s Curve

The Makeham’s curve is of the form

y=a (5.1)
where a , b , c & d are parameters.
Let

, ,
be the values of y corresponding to the values

, ,
of x respectively.
Then the points

( , ),( , ),( , ),( , )

lie on the curve described by equation (1).
Therefore,
_____________________________________________________________________________________
Copy right @ESES IJEAR

64
International Journal of Electronics and Applied Research (IJEAR) vol. 4, issue 1, July 2017
Online (http://eses.net.in/ESES Journal) ISSN 2395-0064

=a , i=0,1,2,3 (5.2)
(5.2) 

log = log a + log b + log c , i=0,1,2,3 (5.3)

If , , are equally spaced then

 =  =  =h

i.e. = +h , = + 2h & = + 3h
Thus,

 log = h log b + (  1) log c , i=0,1,2 (5.4)

& 2 log = (  1)2 log c , i=0,1 (5.5)

Accordingly,

= (5.6)
Thus,

= log ( ) i.e. = log ( )} (5.7)

Thus d can be Obtained from this equation..
Consequently from (5.5),

log c = = (5.8)

i.e. c = antilog { } = antilog { } (5.9)

Thus c can be obtained from this equation where d is given by equation (5.7).
Similarly from (5.4),

log b = { log (  1) log c } , i=0,1,2 (5.10)

which implies,

b= { log (  1) log c }] , i = 0 , 1 , 2 (5.11)

_____________________________________________________________________________________
Copy right @ESES IJEAR

65
International Journal of Electronics and Applied Research (IJEAR) vol. 4, issue 1, July 2017
Online (http://eses.net.in/ESES Journal) ISSN 2395-0064

Thus b can be determined from any one of the three equations given in (5.11) where d & c are given by
equations (5.7) & (5.9) respectively.
Similarly from (5.3),

log a = log  log b  log c , i = 0 , 1 , 2, 3 (5.12)

which implies,

a = antilog {log  log b  log c} , i = 0 , 1 , 2, 3 (5.13)

Thus a can be determined from any one of the four equations given in (5.13) where d, c & b are given by
equations (5.7), (5.9) & (5.11) respectively.

5.1. Example of Data Representation by Makeham’s Curve

Let us represent the data on total populations of India corresponding to the years 1951, 1961, 1971 &
1981 (shown in Table-2.1.1 mentioned in section 2.1) by the modified exponential curve described by
equation (5.1).
As earlier, in order to do it let us take the year 1951 as origin (i.e. 0) and choose the scale 1/10 such
that the values of x corresponding to the years 1961, 1971 & 1981 become 1, 2 & 3 respectively. Now,
let us construct the following table (Table-5.1.1) :
Table-5.1.1

T x P(t) log P(t)  log P(t) log P(t)

1951 0 361088090 19.704632503150 0.195912110819 0.02562041346
1961 1 439234771 19.900544613969 0.221532524279  00112252063
1971 2 548159652 20.122077138248 0.220410003649
1981 3 683329097 20.342487141897

Now by (5.7),
d =  0.043813525170
By (5.9),
c = 1.023793397619

_____________________________________________________________________________________
Copy right @ESES IJEAR

66
International Journal of Electronics and Applied Research (IJEAR) vol. 4, issue 1, July 2017
Online (http://eses.net.in/ESES Journal) ISSN 2395-0064

By (5.11),
b = 1.246646468467
By (5.12),
a = 352696247.934125408029
Thus the Makeham’s curve satisfying the data is

P(t) = a
with the above values of a, b, c & d.
This curve yields,
P(1951) = 352696247.934125408029 × = 361088089 ,
P(1961) = 352696247.934125408029 × ×

= 439234771 ,

P(1971) = 352696247.934125408029 × ×

= 548159652
& P(1981) = 352696247.934125408029 × ×

= 683329097

6. Conclusion

The exponential curve contains two parameters. Accordingly, when two pairs of numerical data are
available, they can be represented by exponential curve. However, when two pairs of numerical data are
available then they can be represented by linear curve also since the linear curve also contains two
parameters. Thus, one question arises -- which of the two curves will suit a set of numerical data better.
The exponential curve described by equation (2.1) satisfies the property that,
 ( log ) = constant
On the other hand, the linear curve described by equation (2.7) satisfies the property that

_____________________________________________________________________________________
Copy right @ESES IJEAR

67
International Journal of Electronics and Applied Research (IJEAR) vol. 4, issue 1, July 2017
Online (http://eses.net.in/ESES Journal) ISSN 2395-0064

 = constant

Thus, comparing the values of   ( log ) one can determine which of the two curves will
suit a set of numerical data better.
The method of representing numerical data on a pair of variables by modified exponential curve
described by equation (3.1) is applicable in representing a set of numerical data on a pair of variables if
the given values of the independent variable are equidistant. In the case where the given values of the
independent variable are equidistant, the method fails to represent the given data by modified exponential
curve. The modified exponential curve contains three parameters. Accordingly, when three pairs of
numerical data are available, they can be represented by modified exponential curve. However, when
three pairs of numerical data are available then they can be represented by quadratic curve also since
quadratic curve also contains three parameters. Thus, one question arises -- which of the two curves will
suit the entire data better.
The quadratic curve is of the form
y = α x2 + β x + γ
where , β , γ are the parameters of the curve
and it satisfies the property that

= constant
On the other hand, the modified exponential curve described by equation (3.1) satisfies the property
that

= constant
Thus, comparing the values of

one can determine which of the two curves will suit a set of numerical data better.
Similarly, the method of representing numerical data on a pair of variables by Pearl & Reed curve
described by equation (4.1) is applicable in representing a set of numerical data on a pair of variables if
the given values of the independent variable are equidistant. In the case where the given values of the
independent variable are not equidistant, the method fails to represent the given data by logistic curve.

_____________________________________________________________________________________
Copy right @ESES IJEAR

68
International Journal of Electronics and Applied Research (IJEAR) vol. 4, issue 1, July 2017
Online (http://eses.net.in/ESES Journal) ISSN 2395-0064

The Pearl & Reed curve contains three parameters. Accordingly, when three pairs of numerical data are
available, they can be represented by logistic curve. However, when three pairs of numerical data are
available then they can also be represented by quadratic curve and by modified exponential curve since
each of these two curves also contains three parameters. As earlier, in this case also one question arises-
which of the three curves will suit a set of numerical data best among the three ones.
The quadratic curve is of the form
y = α x2 + β x + γ
where , β , γ are the parameters of the curve
and it satisfies the property that

= constant
On the other hand, the modified exponential curve described by the equation
y =a + b
where a , b & c are parameters; satisfies the property that

= constant
Again, the Pearl & Reed curve described by the equation (4.1) satisfies the property that

=
Thus, comparing the values of

one can determine which of the three curves will suit the entire data best among the three ones.
Again, the method of representing numerical data on a pair of variables by Makeham’s curve described
by equation (5.1) is applicable in representing a set of numerical data on a pair of variables if the given
values of the independent variable are equidistant. In the case where the given values of the independent
variable are equidistant, the method fails to represent the given data by modified exponential curve. The
Makeham’s curve contains four parameters. Accordingly, when four pairs of numerical data are available,
they can be represented by Makeham’s curve. However, when four pairs of numerical data are available

_____________________________________________________________________________________
Copy right @ESES IJEAR

69
International Journal of Electronics and Applied Research (IJEAR) vol. 4, issue 1, July 2017
Online (http://eses.net.in/ESES Journal) ISSN 2395-0064

then they can be represented by a cubic curve also since cubic curve also contains four parameters. Thus
in this case also, one question arises -- which of the two curves will suit a set of numerical data better.
The cubic curve is of the form
y = α x3 + β x2 + γ x + δ
where , β , γ , δ are the parameters of the curve
and it satisfies the property that

= constant
On the other hand, the modified exponential curve described by equation (3.1) satisfies the property
that

Thus, comparing the values of

one can determine which of the two curves will suit a set of numerical data better.

_____________________________________________________________________________________
Copy right @ESES IJEAR

70
International Journal of Electronics and Applied Research (IJEAR) vol. 4, issue 1, July 2017
Online (http://eses.net.in/ESES Journal) ISSN 2395-0064

References
Abramowitz M., “Handbook of Mathematical Functions with Formulas, Graphs, and Mathematical Tables”, 9th
printing, New York: Dover, 880, 1972.

Abramowitz M & Stegun I. A., "Exponential Function", Handbook of Mathematical Functions with Formulas,
Graphs, and Mathematical Tables, 9th printing, New York: Dover, 69-71, 1972.

Annadurai G., Rajesh Babu S. & Srinivasamoorthy V. R., “Development of mathematical models (Logistic,
Gompertz and Richards models) describing the growth pattern of Pseudomonas putida (NICM 2174)”, Bioprocess
Engineering, Vol. 23, 607-612, 1999.

Bathe K. J. & Wilson E. L., “Numerical Methods in Finite Element Analysis”, Prentice-Hall, Englewood Cliffs,
NJ, 1976.

Beyer W. H., “CRC Standard Mathematical Tables, 28th ed.” Boca Raton, FL: CRC Press, pp. 217, 1987.
Borwein J., Bailey D. & Girgensohn R., “Experimentation in Mathematics: Computational Paths to Discovery”.
Wellesley, MA: A K Peters, 2004.

Chakrabarty Dhritikesh, “Representation of Numerical Data on a Pair of Variables by a Polynomial Curve

Expressed in the Simplest Form”, International Journal of Electronics and Applied Research, Vol. 3, No. 1, 26 –
39, 2016a.

Chakrabarty Dhritikesh, “Interpolation: One Method of Representation of Numerical Data on a Pair of Variables by
a Polynomial Curve Expressed In the Simplest Form”, Journal of Environmental Science, Computer Science and
Engineering & Technology, Section C, Vol. 5, No. 3, 405 – 418, 2016b.

Chakrabarty Dhritikesh, “Difference and Ratio Operators: Representation of Numerical Data on a Pair of Variables
by a Polynomial Curve”, Journal of Environmental Science, Computer Science and engineering & Technology,
Section C, Vol. 5, No. 4, 549 –560, 2016c.

Chakrabarty Dhritikesh, “Recent Developments on Representation of Numerical Data by a Polynomial Curve”,

International Journal of Electronics and Applied Research, Vol. 3, No. 2, 125 – 158, 2016d.

Chakrabarty Dhritikesh, “Backward Difference and Ratio Operations: Representation of Numerical Data on a Pair
of Variables by a Polynomial Curve”, Journal of Environmental Science, Computer Science and Engineering &
Technology, Section C, Vol. 6, No. 1,027 –039, 2016 – 17, Also available in www.jecet.org .

Chakrabarty Dhritikesh, “Some Forms of Interpolation Formula Based on Divided Difference”, Journal of
Environmental Science, Computer Science and Engineering & Technology, Section C, Vol. 6, No. 2, 199 – 211,
2017, Also available in www.jecet.org .

Cayley A., "A Memoir on the Theory of Matrices". Phil.Trans., Vol. 148, 1858.

Cayley A., “The Collected Mathematical Papers of Arthur Cayley (Classic Reprint)”, Forgotten books, ASIN
B008HUED9O, 1889.

Chapra S. C. & Canale R. P. , “Numerical Methods for Engineers”, third ed., McGraw-Hill, NewYork, 2002.

_____________________________________________________________________________________
Copy right @ESES IJEAR

71
International Journal of Electronics and Applied Research (IJEAR) vol. 4, issue 1, July 2017
Online (http://eses.net.in/ESES Journal) ISSN 2395-0064

Chwaiger J.,“On a Characterization of Polynomials by Divided Differences”, Aequationes Math, Vol. 48, 317 –
323, 1994.

Conte S. D. & Boor Carl de, “Elementary Numerical Analysis”, 3 Ed, McGraw-Hill, New York, USA, 1980.

Corliss J. J., “Note on an Extension of Lagrange's Formula”, American Mathematical Monthly Jstor, Vol. 45, No.
2, 106 – 107, 1938.

Das Biswajit & Chakrabarty Dhritikesh, “Lagranges Interpolation Formula: Representation of Numerical Data by a
Polynomial Curve”, International Journal of Mathematics Trend and Technology Vol. 34 (part-1), No. 2, 23 – 31,
2016a.

Das Biswajit & Chakrabarty Dhritikesh, “Newton’s Divided Difference Interpolation Formula: Representation of
Numerical Data by a Polynomial Curve”, International Journal of Mathematics Trend and Technology, Vol. 35
(part-1), No. 3, 26 – 32, 2016b.

Das Biswajit & Chakrabarty Dhritikesh, “Newton’s Forward Interpolation Formula: Representation of Numerical
Data by a Polynomial Curve”, International Journal of Applied Research, Vol. 1, No. 2, 36 – 41, 2016c.

Das Biswajit & Chakrabarty Dhritikesh, “Newton’s Backward Interpolation Formula: Representation of Numerical
Data by a Polynomial Curve”, International Journal of Statistics and Applied Mathematics, Vol.2, No. 10,513 –
517, 2016d.

Das Biswajit & Chakrabarty Dhritikesh, “Matrix Inversion: Representation of Numerical Data by a Polynomial
curve”, Aryabhatta Journal of Mathematics & Informatics, Vol. 8, No. 2, pp. 267 – 276, 2016e.

Das Biswajit & Chakrabarty Dhritikesh, “Inversion of Matrix by Elementary Transformation: Representation of
Numerical Data by a Polynomial Curve”, Journal of Mathematis and System Sciences, Vol. 12, No., 27 – 32,
2016f.

Das Biswajit & Chakrabarty Dhritikesh, “Inversion of Matrix by Elementary Column Transformation:
Representation of Numerical Data by a Polynomial Curve”, International Journal of Mathematics Trends and
Technology, Vol. 42, No. 1, 45 – 49, 2017a, Also available in http://www.ijmttjournal.org .

Das Biswajit & Chakrabarty Dhritikesh, “Backward Divided Difference: Representation of Numerical Data by a
Polynomial Curve”, International Journal of Statistics and Applied Mathematics, Vol. 2, No. 2, 1 – 6, 2017b, Also
available in http://www.mathsjournal.com .

De Boor C., “A divided difference expansion of a divided difference”, J. Approx. Theory, Vol. 122, 10 – 12, 2003.

Dokken T. & Lyche T., “A divided difference formula for the error in Hermite interpolation”, BIT, Vol. 19, 540 –
541, 1979.

Erdos P. & Turan P., “On Interpolation II: On the Distribution of the Fundamental Points of Lagrange and Hermite
Interpolation”, The Annals of Mathematics, 2nd Ser”, Jstor, Vol. 39, No. 4, 703 – 724, 1938.
Echols W. H., “On Some Forms of Lagrange's Interpolation Formula”, “The Annals of Mathematics, Jstor, Vol. 8,
No. 1, 22 – 24, 1893.

_____________________________________________________________________________________
Copy right @ESES IJEAR

72
International Journal of Electronics and Applied Research (IJEAR) vol. 4, issue 1, July 2017
Online (http://eses.net.in/ESES Journal) ISSN 2395-0064

Endre S. & David M. , “An Introduction to Numerical Analysis”, Cambridge, UK, 2003.

Floater M. , “Error formulas for divided difference expansions and numerical differentiation”, J. Approx. Theory,
Vol. 122, 1 – 9, 2003.

Fred T., “Recurrence Relations for Computing with Modified Divided Differences”, Mathematics of Computation,
Vol. 33, No. 148, 1265 – 1271, 1979.

Gertrude B. , “On Modified Divided Differences”, Mathematical Tables and Other Aids to Computation, Vol. 8,
No. 45, 1 – 11, 1954.

Gerald C. F. & Wheatley P.O., “Applied Numerical Analysis”, fifth ed., Addison-Wesley Pub. Co., MA, 1994.

Gershenfeld Neil A. “The Nature of Mathematical Modeling”,Cambridge, UK: Cambridge University Press, 1999.

Herbert E. S., “Multi-Point Generalization of Newton's Divided Difference Formula”, Proceedings of the American
Mathematical Society, Jstor, Vol. 13, No. 2, 210 – 212, 1962.

Hummel P. M., “A Note on Interpolation (in Mathematical Notes)”, American Mathematical Monthly, Jstor, Vol.
54, No. 4, 218 – 219, 1947.

Hamilton W. R., “Lectures on Quaternions”, Dublin , 1853.

Hamilton W. R., “On the Existence of a Symbolic and Biquadratic Equation which is satisfied by the Symbol of
Linear or Distributive Operation on a Quaternion”, The London, Edinburgh and Dublin Philosophical Magazine
and Journal of Science, Series IV, Taylor & Francis 24, 127 – 128, 1862.

Hamilton W. R., “On a New and General Method of Inverting a Linear and Quaternion Function of a Quaternion”,
Proceedings of the Royal Irish Academy, Royal Irish Academy, Vol. 8, 182 – 183, 1864a.

Hamilton W. R. , “On the Existence of a Symbolic and Biquadratic Equation, which is satisfied by the Symbol of
Linear Operation in Quaternions”, Proceedings of the Royal Irish Academy, Royal Irish Academy, Vol. 8, 190–101,
1864b.

Jannedy Stefanie, Bod Rens & Hay Jennifer., “Probabilistic Linguistics”, Cambridge, Massachusetts: MIT Press,
2003.

Jan K. W., “Note on Interpolation (in Notes)”, Journal of the American Statistical Association, Jstor, Vol. 25, 170,
203 – 205, 1930.

Jeffreys H. & Jeffreys B. S., “Divided Differences” , Methods of Mathematical Physics, 3rd ed, 260 – 264, 1988.

John H. M. & Kurtis D.F. , “Numerical methods using MATLAB”, 4ed, Pearson Education, USA, 2004.

Jordan C., “Calculus of Finite Differences”, 3rd ed, New York, Chelsea, 1965.

Kendall E. Atkinson , “An Introduction to Numerical Analysis”, 2nd Ed., New York, 1989.

_____________________________________________________________________________________
Copy right @ESES IJEAR

73
International Journal of Electronics and Applied Research (IJEAR) vol. 4, issue 1, July 2017
Online (http://eses.net.in/ESES Journal) ISSN 2395-0064

Kincard David R. & Ward Chaney E., “Numerical analysis”, Brooks /Cole, Pacific Grove, CA, 1991.

Kingsland Sharon E. “Modeling nature: episodes in the history of population ecology”. Chicago: University of
Chicago Press, 1995.

Krantz S. G., "The Exponential and Applications.", Handbook of Complex Variables. Boston, MA: Birkhäuser,
pp. 7-12, 1999.

Lee E. T. Y. , “A Remark on Divided Differences”, American Mathematical Monthly, Vol. 96, No 7, 618 – 622,
1989.

Mills T. M. , “An introduction to analysis by Lagrange interpolation”, Austral. Math. Soc. Gaz., MathSciNet, Vol.
4, No. 1, 10 – 18, 1977.

Makeham W. M., "On the Law of Mortality and the Construction of Annuity Tables." J. Inst. Actuaries and Assur.
Mag., Vol. 8, 301-310, 1860.

Makeham W. M., "On an Application of the Theory of the Composition of Decremental Forces." J. Inst. Actuaries
and Assur. Mag., Vol. 18, 317-322, 1874.

Nasrin Akter Ripa, “Analysis of Newton’s Forward Interpolation Formula”, International Journal of Computer
Science & Emerging Technologies, Vol. 1, No. 4, 2010.

Neale E. P. & Sommerville D. M. Y. , “A Shortened Interpolation Formula for Certain Types of Data”, Journal of
the American Statistical Association, Jstor, Vol. 19, No. 148, 515 – 517, 1924.

Quadling D. A. , “Lagrange's Interpolation Formula”, The Mathematical Gazette, Vol. L, No. 374, pp. 372 – 375,
1966.

Robert J. Schilling & Sandra L. Harries, “Applied Numerical Methods for Engineers”, Brooks /Cole, Pacific
Grove, CA, 2000.

Revers & Michael Bull, “On Lagrange interpolation with equally spaced nodes”, Austral. Math. Soc , MathSciNet.,
Vol. 62, No. 3, 357 – 368, 2000.

Scarborough James B., “Numerical Mathematical Analysis”, 6 Ed, The John Hopkins Press, USA, 1996.
Traub J. F., “On Lagrange-Hermite Interpolation”, Journal of the Society for Industrial and Applied Mathematics,
Jstor, Vol. 12, No. 4, 886 – 891, 1964.

Vertesi P., “SIAM Journal on Numerical Analysis”, Jstor , Vol. 27, No. 5, 1322-1331, 1990.

Whittaker E. T. and Robinson G., “The Gregory-Newton Formula of Interpolation" and "An Alternative Form of
the Gregory-Newton Formula", The Calculus of Observations: A Treatise on Numerical Mathematics, 4th ed, 10 –
15, 1967a.

Whittaker E. T. & Robinson G., “Divided Differences & Theorems on Divided Differences”, The Calculus of
Observations: A Treatise on Numerical Mathematics, 4th ed., New York, 20 – 24, 1967b.

_____________________________________________________________________________________
Copy right @ESES IJEAR

74

View publication stats