Backward divided difference: Representation of numerical data by a polynomial curve

Introduction

There exist a number of interpolation formulae such as 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 etc. in the literature of numerical analysis {Bathe & Wilson (1976), Jan (1930), Hummel (1947) et al} for handling the problems of interpolation {Hummel (1947), Erdos & Turan (1938) et al}. In the existing approach of interpolation {Hummel (1947), Erdos & Turan (1938) et al}, where a number of interpolation formulae are available {Bathe & Wilson (1976), Jan (1930), Hummel (1947) et al}, 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 then 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 is necessary for representing a given set of numerical data on a pair of variables by a suitable polynomial. Das & Chakrabarty (2016a, 2016b, 2016c & 2016d) [2-5] derived four formulae for representing numerical data on a pair of variables by a polynomial curve. They derived the formulae from Lagranges Interpolation Formula, Newton's Divided Difference Interpolation Formula, Newton's Forward Interpolation Formula and Newton's Backward Interpolation respectively. In this study, another formula has been derived from Newton's Divided Difference formula applied in reverse order. This paper describes the derivation of the formula with numerical example in order to show the application of the formula to numerical data.

Backward Divided Difference Interpolation Formula:

Let y f x , y f x , y f x , …………, y  f x  , y f x be the values of the function y = f(x) corresponding to the values x , x , x , …………, x  , x of the independent variable x. Then Newton's Divided Difference uses interpolation formula is given by

x for n = 1, 2, 3, ………….., n. Now, applying the formula to the above values of x & f(x) arranged in reverse order i.e. applying the formula to the values

Representation of Numerical Data by Polynomial Curve:

By algebraic expansion, one can obtain that

x x ) Now, Backward divided difference interpolation formula, described by equation ( 1) can be expressed as f

Coefficient of x n = C0 The equation ( 2), can be expressed as f(x) = An + An  1 x + An  2 x 2 + An  3 x 3 + …………………….. + A0 x n (3) which is the required formula for representation of numerical data by a polynomial curve where An = Cn Cn 3), with the coefficients An, An  1, An  2, An  3, …………………….., A0 ,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.

Example of Application of the Formula: Example (4.1):

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

= 171542 + 82012.16 (0 +1 +2) + 66584.99 (0 + 0

Thus, the polynomial that can represent the given numerical data is (4)  f(x) = 31205576  5302245.

= 2545558.66 + 217007.25 (0 + 1 + 2 + 3) = 3847602.16 A0 = C0 = 217007.25 Thus, the polynomial that can represent the given numerical data is (5)  f(x) = 1210193422 178017822.18 x  8790947.73x

Conclusion

The formula described by equation ( 3) can be used to represent a given set of numerical data on a pair of variables, by a polynomial.

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. Newton's forward interpolation formula is valid for estimating the value of the dependent variable under the following two conditions: (i) The given values of the independent variable are at equal interval.

(ii) The value of the independent variable corresponding to which the value of the dependent variable is to be estimated lies in the first half of the series of the given values of the independent variable.

However, Newton's divided difference interpolation formula is valid for estimating the value of the dependent variable beyond these two conditions. Therefore, the formula derived here is valid for representing a set of numerical data on a pair of variables by a polynomial beyond these two conditions.

References

1. Chwaiger J. On a Characterization of Polynomials by Divided Differences", "Aequationes Math, 1994; 48:317-323.
2. De Boor C. A divided difference expansion of a divided difference, J. Approx. Theory, 2003; 122:10-12.
3. Dokken T, Lyche T. A divided difference formula for the error in Hermite interpolation, BIT, 1979; 19:540-41.
4. Fred T, Recurrence Relations for Computing with Modified Divided Differences, Mathematics of Computation, 1979; 33(148):1265-1271.
5. Floater M. Error formulas for divided difference expansions and numerical differentiation, J. Approx. Theory, 2003; 122:1-9.
6. Gertrude Blanch. On Modified Divided Differences, Mathematical Tables and Other Aids to Computation, 1954; 8(45):1-11.
7. Jeffreys H, Jeffreys BS. Divided Differences, Methods of Mathematical Physics, 3 rd ed, 1988, 260-264, 14. Lee ETY. A Remark on Divided Differences, American Mathematical Monthly, 1989; 96(7):618-622.
8. Whittaker ET, Robinson G. Divided Differences & Theorems on Divided Differences, The Calculus of Observations: A Treatise on Numerical Mathematics", 4 th ed., New York, 1967, 20-24.
9. Wang X, Yang S. On divided differences of the remainder of polynomial interpolation, 2004. www.math.uga.edu/~mjlai/pub.html.
10. Wang X, Wang H. Some results on numerical divided difference formulas, 2003. www.math.uga.edu/~mjlai/pub.html.