Divided differences

In mathematics, divided differences is an algorithm, historically used for computing tables of logarithms and trigonometric functions.^{[citation needed]} Charles Babbage's difference engine, an early mechanical calculator, was designed to use this algorithm in its operation.

Divided differences is a recursive division process. Given a sequence of data points $(x_{0}, y_{0}), \dots, (x_{n}, y_{n})$ , the method calculates the coefficients of the interpolation polynomial of these points in the Newton form.

It is sometimes denoted by a delta with a bar: $△ |$ or $◿ ◺$ .

Definition

Given n + 1 data points $(x_{0}, y_{0}), \dots, (x_{n}, y_{n})$ where the $x_{k}$ are assumed to be pairwise distinct, the forward divided differences are defined as: $\begin{matrix} [y_{k}] & := y_{k}, & k \in {0, \dots, n} \\ [y_{k}, \dots, y_{j}] & := \frac{[y_{k + 1}, \dots, y_{j}] - [y_{k}, \dots, y_{j - 1}]}{x_{j} - x_{k}}, & k \in {0, \dots, n - 1}, j \in {k + 1, \dots, n} . \end{matrix}$

To make the recursive process of computation clearer, the divided differences can be put in tabular form, where the columns correspond to the value of j above, and each entry in the table is computed from the difference of the entries to its immediate lower left and to its immediate upper left, divided by a difference of corresponding x-values: $\begin{matrix} x_{0} & y_{0} = [y_{0}] \\ [y_{0}, y_{1}] \\ x_{1} & y_{1} = [y_{1}] & [y_{0}, y_{1}, y_{2}] \\ [y_{1}, y_{2}] & [y_{0}, y_{1}, y_{2}, y_{3}] \\ x_{2} & y_{2} = [y_{2}] & [y_{1}, y_{2}, y_{3}] \\ [y_{2}, y_{3}] \\ x_{3} & y_{3} = [y_{3}] \end{matrix}$

Notation

Note that the divided difference $[y_{k}, \dots, y_{k + j}]$ depends on the values $x_{k}, \dots, x_{k + j}$ and $y_{k}, \dots, y_{k + j}$ , but the notation hides the dependency on the x-values. If the data points are given by a function f, $(x_{0}, y_{0}), \dots, (x_{n}, y_{n}) = (x_{0}, f (x_{0})), \dots, (x_{n}, f (x_{n}))$ one sometimes writes the divided difference in the notation $f [x_{k}, \dots, x_{k + j}] \overset{def}{=} [f (x_{k}), \dots, f (x_{k + j})] = [y_{k}, \dots, y_{k + j}] .$ Other notations for the divided difference of the function ƒ on the nodes x₀, ..., x_n are: $f [x_{k}, \dots, x_{k + j}] = [x_{0}, \dots, x_{n}] f = [x_{0}, \dots, x_{n}; f] = D [x_{0}, \dots, x_{n}] f .$

Example

Divided differences for $k = 0$ and the first few values of $j$ : $\begin{matrix} [y_{0}] & = y_{0} \\ [y_{0}, y_{1}] & = \frac{y_{1} - y_{0}}{x_{1} - x_{0}} \\ [y_{0}, y_{1}, y_{2}] & = \frac{[y_{1}, y_{2}] - [y_{0}, y_{1}]}{x_{2} - x_{0}} = \frac{\frac{y_{2} - y_{1}}{x_{2} - x_{1}} - \frac{y_{1} - y_{0}}{x_{1} - x_{0}}}{x_{2} - x_{0}} = \frac{y_{2} - y_{1}}{(x_{2} - x_{1}) (x_{2} - x_{0})} - \frac{y_{1} - y_{0}}{(x_{1} - x_{0}) (x_{2} - x_{0})} \\ [y_{0}, y_{1}, y_{2}, y_{3}] & = \frac{[y_{1}, y_{2}, y_{3}] - [y_{0}, y_{1}, y_{2}]}{x_{3} - x_{0}} \end{matrix}$

Thus, the table corresponding to these terms up to two columns has the following form: $\begin{matrix} x_{0} & y_{0} \\ \frac{y_{1} - y_{0}}{x_{1} - x_{0}} \\ x_{1} & y_{1} & \frac{\frac{y_{2} - y_{1}}{x_{2} - x_{1}} - \frac{y_{1} - y_{0}}{x_{1} - x_{0}}}{x_{2} - x_{0}} \\ \frac{y_{2} - y_{1}}{x_{2} - x_{1}} \\ x_{2} & y_{2} & ⋮ \\ ⋮ \\ ⋮ & ⋮ \\ ⋮ \\ x_{n} & y_{n} \end{matrix}$

Properties

Linearity $\begin{matrix} (f + g) [x_{0}, \dots, x_{n}] & = f [x_{0}, \dots, x_{n}] + g [x_{0}, \dots, x_{n}] \\ (λ \cdot f) [x_{0}, \dots, x_{n}] & = λ \cdot f [x_{0}, \dots, x_{n}] \end{matrix}$
Leibniz rule $(f \cdot g) [x_{0}, \dots, x_{n}] = f [x_{0}] \cdot g [x_{0}, \dots, x_{n}] + f [x_{0}, x_{1}] \cdot g [x_{1}, \dots, x_{n}] + \dots + f [x_{0}, \dots, x_{n}] \cdot g [x_{n}] = \sum_{r = 0}^{n} f [x_{0}, \dots, x_{r}] \cdot g [x_{r}, \dots, x_{n}]$
Divided differences are symmetric: If $σ : {0, \dots, n} \to {0, \dots, n}$ is a permutation then $f [x_{0}, \dots, x_{n}] = f [x_{σ (0)}, \dots, x_{σ (n)}]$
Polynomial interpolation in the Newton form: if $P$ is a polynomial function of degree $\leq n$ , and $p [x_{0}, \dots, x_{n}]$ is the divided difference, then $P_{n - 1} (x) = p [x_{0}] + p [x_{0}, x_{1}] (x - x_{0}) + p [x_{0}, x_{1}, x_{2}] (x - x_{0}) (x - x_{1}) + \dots + p [x_{0}, \dots, x_{n}] (x - x_{0}) (x - x_{1}) \dots (x - x_{n - 1})$
If $p$ is a polynomial function of degree $< n$ , then $p [x_{0}, \dots, x_{n}] = 0.$
Mean value theorem for divided differences: if $f$ is n times differentiable, then $f [x_{0}, \dots, x_{n}] = \frac{f^{(n)} (ξ)}{n!}$ for a number $ξ$ in the open interval determined by the smallest and largest of the $x_{k}$ 's.

Matrix form

The divided difference scheme can be put into an upper triangular matrix: $T_{f} (x_{0}, \dots, x_{n}) = (\begin{matrix} f [x_{0}] & f [x_{0}, x_{1}] & f [x_{0}, x_{1}, x_{2}] & \dots & f [x_{0}, \dots, x_{n}] \\ 0 & f [x_{1}] & f [x_{1}, x_{2}] & \dots & f [x_{1}, \dots, x_{n}] \\ 0 & 0 & f [x_{2}] & \dots & f [x_{2}, \dots, x_{n}] \\ ⋮ & ⋮ & ⋱ & ⋮ \\ 0 & 0 & 0 & \dots & f [x_{n}] \end{matrix}) .$

Then it holds

$T_{f + g} (x) = T_{f} (x) + T_{g} (x)$
$T_{λ f} (x) = λ T_{f} (x)$ if $λ$ is a scalar
$T_{f \cdot g} (x) = T_{f} (x) \cdot T_{g} (x)$
This follows from the Leibniz rule. It means that multiplication of such matrices is commutative. Summarised, the matrices of divided difference schemes with respect to the same set of nodes x form a commutative ring.
Since $T_{f} (x)$ is a triangular matrix, its eigenvalues are obviously $f (x_{0}), \dots, f (x_{n})$ .
Let $δ_{ξ}$ be a Kronecker delta-like function, that is $δ_{ξ} (t) = {\begin{matrix} 1 & : t = ξ, \\ 0 & : else . \end{matrix}$ Obviously $f \cdot δ_{ξ} = f (ξ) \cdot δ_{ξ}$ , thus $δ_{ξ}$ is an eigenfunction of the pointwise function multiplication. That is $T_{δ_{x_{i}}} (x)$ is somehow an "eigenmatrix" of $T_{f} (x)$ : $T_{f} (x) \cdot T_{δ_{x_{i}}} (x) = f (x_{i}) \cdot T_{δ_{x_{i}}} (x)$ . However, all columns of $T_{δ_{x_{i}}} (x)$ are multiples of each other, the matrix rank of $T_{δ_{x_{i}}} (x)$ is 1. So you can compose the matrix of all eigenvectors of $T_{f} (x)$ from the $i$ -th column of each $T_{δ_{x_{i}}} (x)$ . Denote the matrix of eigenvectors with $U (x)$ . Example $U (x_{0}, x_{1}, x_{2}, x_{3}) = (\begin{matrix} 1 & \frac{1}{(x_{1} - x_{0})} & \frac{1}{(x_{2} - x_{0}) (x_{2} - x_{1})} & \frac{1}{(x_{3} - x_{0}) (x_{3} - x_{1}) (x_{3} - x_{2})} \\ 0 & 1 & \frac{1}{(x_{2} - x_{1})} & \frac{1}{(x_{3} - x_{1}) (x_{3} - x_{2})} \\ 0 & 0 & 1 & \frac{1}{(x_{3} - x_{2})} \\ 0 & 0 & 0 & 1 \end{matrix})$ The diagonalization of $T_{f} (x)$ can be written as $U (x) \cdot diag (f (x_{0}), \dots, f (x_{n})) = T_{f} (x) \cdot U (x) .$

Polynomials and power series

The matrix $J = (\begin{matrix} x_{0} & 1 & 0 & 0 & \dots & 0 \\ 0 & x_{1} & 1 & 0 & \dots & 0 \\ 0 & 0 & x_{2} & 1 & 0 \\ ⋮ & ⋮ & ⋱ & ⋱ \\ 0 & 0 & 0 & 0 & ⋱ & 1 \\ 0 & 0 & 0 & 0 & x_{n} \end{matrix})$ contains the divided difference scheme for the identity function with respect to the nodes $x_{0}, \dots, x_{n}$ , thus $J^{m}$ contains the divided differences for the power function with exponent $m$ . Consequently, you can obtain the divided differences for a polynomial function $p$ by applying $p$ to the matrix $J$ : If $p (ξ) = a_{0} + a_{1} \cdot ξ + \dots + a_{m} \cdot ξ^{m}$ and $p (J) = a_{0} + a_{1} \cdot J + \dots + a_{m} \cdot J^{m}$ then $T_{p} (x) = p (J) .$ This is known as Opitz' formula.

Now consider increasing the degree of $p$ to infinity, i.e. turn the Taylor polynomial into a Taylor series. Let $f$ be a function which corresponds to a power series. You can compute the divided difference scheme for $f$ by applying the corresponding matrix series to $J$ : If $f (ξ) = \sum_{k = 0}^{\infty} a_{k} ξ^{k}$ and $f (J) = \sum_{k = 0}^{\infty} a_{k} J^{k}$ then $T_{f} (x) = f (J) .$

Alternative characterizations

Expanded form

$\begin{matrix} f [x_{0}] & = f (x_{0}) \\ f [x_{0}, x_{1}] & = \frac{f (x_{0})}{(x_{0} - x_{1})} + \frac{f (x_{1})}{(x_{1} - x_{0})} \\ f [x_{0}, x_{1}, x_{2}] & = \frac{f (x_{0})}{(x_{0} - x_{1}) \cdot (x_{0} - x_{2})} + \frac{f (x_{1})}{(x_{1} - x_{0}) \cdot (x_{1} - x_{2})} + \frac{f (x_{2})}{(x_{2} - x_{0}) \cdot (x_{2} - x_{1})} \\ f [x_{0}, x_{1}, x_{2}, x_{3}] & = \frac{f (x_{0})}{(x_{0} - x_{1}) \cdot (x_{0} - x_{2}) \cdot (x_{0} - x_{3})} + \frac{f (x_{1})}{(x_{1} - x_{0}) \cdot (x_{1} - x_{2}) \cdot (x_{1} - x_{3})} + \\ \frac{f (x_{2})}{(x_{2} - x_{0}) \cdot (x_{2} - x_{1}) \cdot (x_{2} - x_{3})} + \frac{f (x_{3})}{(x_{3} - x_{0}) \cdot (x_{3} - x_{1}) \cdot (x_{3} - x_{2})} \\ f [x_{0}, \dots, x_{n}] & = \sum_{j = 0}^{n} \frac{f (x_{j})}{\prod_{k \in {0, \dots, n} ∖ {j}} (x_{j} - x_{k})} \end{matrix}$

With the help of the polynomial function $ω (ξ) = (ξ - x_{0}) \dots (ξ - x_{n})$ this can be written as $f [x_{0}, \dots, x_{n}] = \sum_{j = 0}^{n} \frac{f (x_{j})}{ω^{'} (x_{j})} .$

Peano form

If $x_{0} < x_{1} < \dots < x_{n}$ and $n \geq 1$ , the divided differences can be expressed as $f [x_{0}, \dots, x_{n}] = \frac{1}{(n - 1)!} \int_{x_{0}}^{x_{n}} f^{(n)} (t) B_{n - 1} (t) d t$ where $f^{(n)}$ is the $n$ -th derivative of the function $f$ and $B_{n - 1}$ is a certain B-spline of degree $n - 1$ for the data points $x_{0}, \dots, x_{n}$ , given by the formula $B_{n - 1} (t) = \sum_{k = 0}^{n} \frac{(max (0, x_{k} - t))^{n - 1}}{ω^{'} (x_{k})}$

This is a consequence of the Peano kernel theorem; it is called the Peano form of the divided differences and $B_{n - 1}$ is the Peano kernel for the divided differences, all named after Giuseppe Peano.

Hermite–Genocchi formula

When ⁠ $f$ ⁠ is ⁠ $n$ ⁠ times continuously differentiable on a convex domain containing the nodes ⁠ $x_{0}, \dots, x_{n}$ ⁠, including the confluent case of repeated nodes in arbitrary order (see § Generalization), the divided difference can be expressed as an integral over a ⁠ $n$ ⁠-simplex ⁠ $Δ_{n}$ ⁠: $f [x_{0}, \dots, x_{n}] = \int_{Δ_{n}} f^{(n)} (\sum_{k = 0}^{n} t_{k} x_{k}) d t_{0} \dots d t_{n},$ where $Δ_{n} = {(t_{0}, \dots, t_{n}) \in ℝ^{n} ∣ t_{i} \geq 0, \sum_{k = 0}^{n} t_{k} = 1}$ is the ⁠ $n$ ⁠-simplex.

Cauchy's integral formula

When ⁠ $f$ ⁠ is analytic on a simply connected domain ⁠ $Ω \subset ℂ$ ⁠ containing the nodes ⁠ $x_{0}, \dots, x_{n}$ ⁠, including the confluent case of repeated nodes in arbitrary order (see § Generalization), the divided difference admits a contour integral representation via Cauchy's integral formula: $f [x_{0}, \dots, x_{n}] = \frac{1}{2 π i} \oint_{Γ} \frac{f (z)}{\prod_{k = 0}^{n} (z - x_{k})} d z,$ where ⁠ $Γ$ ⁠ is a positively oriented, simple closed contour in ⁠ $Ω$ ⁠ enclosing all the nodes ⁠ $x_{0}, \dots, x_{n}$ ⁠.

Forward and backward differences

When the data points are equidistantly distributed we get the special case called forward differences. They are easier to calculate than the more general divided differences.

Given n+1 data points $(x_{0}, y_{0}), \dots, (x_{n}, y_{n})$ with $x_{k} = x_{0} + k h, for k = 0, \dots, n and fixed h > 0$ the forward differences are defined as: $\begin{matrix} Δ^{(0)} y_{k} & := y_{k}, k = 0, \dots, n \\ Δ^{(j)} y_{k} & := Δ^{(j - 1)} y_{k + 1} - Δ^{(j - 1)} y_{k}, k = 0, \dots, n - j, j = 1, \dots, n . \end{matrix}$
whereas the backward differences are defined as: $\begin{matrix} \nabla^{(0)} y_{k} & := y_{k}, k = 0, \dots, n \\ \nabla^{(j)} y_{k} & := \nabla^{(j - 1)} y_{k} - \nabla^{(j - 1)} y_{k - 1}, k = 0, \dots, n - j, j = 1, \dots, n . \end{matrix}$
Thus the forward difference table is written as: $\begin{matrix} y_{0} \\ Δ y_{0} \\ y_{1} & Δ^{2} y_{0} \\ Δ y_{1} & Δ^{3} y_{0} \\ y_{2} & Δ^{2} y_{1} \\ Δ y_{2} \\ y_{3} \end{matrix}$
whereas the backwards difference table is written as: $\begin{matrix} y_{0} \\ \nabla y_{1} \\ y_{1} & \nabla^{2} y_{2} \\ \nabla y_{2} & \nabla^{3} y_{3} \\ y_{2} & \nabla^{2} y_{3} \\ \nabla y_{3} \\ y_{3} \end{matrix}$

The relationship between divided differences and forward differences is $[y_{j}, y_{j + 1}, \dots, y_{j + k}] = \frac{1}{k! h^{k}} Δ^{(k)} y_{j},$ whereas for backward differences:^{[citation needed]} $[y_{j}, y_{j - 1}, \dots, y_{j - k}] = \frac{1}{k! h^{k}} \nabla^{(k)} y_{j} .$

Explicit formula

When the data points are equispaced we can also derive an explicit formula for $[y_{j}, \dots, y_{j + k}]$ . For any fixed $k \leq n$ and $j$ such that $j + k \leq n$ , $[y_{j}, \dots, y_{j + k}] = \frac{(- 1)^{j + k}}{k! h^{k}} \sum_{i = j}^{j + k} (- 1)^{i} y_{i} (\binom{k}{i - j}) .$

Proof. We prove this by induction on $k$ :

Base case: Let $k = 0$ and $j = 0 \dots n$ . Then by definition $[y_{j}] = y_{j}$ , and $\frac{(- 1)^{j}}{0! h^{0}} \sum_{i = j}^{j} (- 1)^{i} y_{i} (\binom{0}{i - j}) = (- 1)^{2 j} y_{j} = y_{j} .$

Induction step: Assume the above formula holds until $k$ and consider $j$ such that $j + k + 1 \leq n$ . By the recursive definition we have

$[y_{j}, \dots, y_{j + k + 1}] = \frac{1}{h (k + 1)} ([y_{j + 1}, \dots, y_{j + k + 1}] - [y_{j}, \dots, y_{j + k}]) .$

We can use our inductive hypothesis on both members of the left side, and obtain

$\frac{1}{h (k + 1)} (\frac{(- 1)^{j + k + 1}}{k! h^{k}} \sum_{i = j + 1}^{j + k + 1} (- 1)^{i} y_{i} (\binom{k}{i - j}) - \frac{(- 1)^{j + k}}{k! h^{k}} \sum_{i = j}^{j + k} (- 1)^{i} y_{i} (\binom{k}{i - j})) .$

This can be rearranged as

$\frac{(- 1)^{j + k + 1}}{(k + 1)! h^{k + 1}} \sum_{i = j}^{j + k + 1} (- 1)^{i} y_{i} ((\binom{k}{i - j}) + (\binom{k}{i - j - 1}))$

where $(\binom{k}{s}) = 0$ when $s < k$ . We obtain our thesis for $k + 1$ by substituting the identity $(\binom{k}{i - j}) + (\binom{k}{i - j - 1}) = (\binom{k + 1}{i - j}) .$

Generalization

When some of the nodes ⁠ $x_{0}, \dots, x_{n}$ ⁠ coincide, the recursive definition of standard divided differences involves division by zero and can no longer be applied directly. In this confluent case, divided differences incorporate higher-order derivatives of the function ⁠ $f$ ⁠. The resulting values are known as confluent divided differences or generalized divided differences. Conceptually, node confluence bridges discrete divided differences and continuous calculus, replacing finite differences with exact higher-order derivatives as nodes coalesce.

Let ⁠ $x_{0}, \dots, x_{n}$ ⁠ consist of nodes with arbitrary repetitions and order. Suppose there are ⁠ $m$ ⁠ distinct values ⁠ $z_{1}, \dots, z_{m}$ ⁠ among the nodes, with corresponding multiplicities ⁠ $μ_{1}, \dots, μ_{m} \geq 1$ ⁠ such that ⁠ $\sum_{i = 1}^{m} μ_{i} = n + 1$ ⁠. Let ⁠ $f$ ⁠ be a function that is ⁠ $μ_{i} - 1$ ⁠ times differentiable at ⁠ $z_{i}$ ⁠ for ⁠ $i = 1, \dots, m$ ⁠.

Confluent divided differences can be defined as the limit: $f [x_{0}, \dots, x_{n}] = lim_{ξ_{0} \to x_{0}, \dots, ξ_{n} \to x_{n}} f [ξ_{0}, \dots, ξ_{n}]$ where ⁠ $ξ_{0}, \dots, ξ_{n}$ ⁠ are pairwise distinct nodes converging to ⁠ $x_{0}, \dots, x_{n}$ ⁠, provided the limit exists.

In practice, any confluent divided difference with arbitrarily ordered repeated nodes can be computed by first permuting the node sequence so that identical nodes are adjacent, and then constructing the divided difference table with an additional higher-order derivative substitution rule.

Divided differences remain symmetric in ⁠ $x_{0}, \dots, x_{n}$ ⁠ in the confluent case, satisfying: $f [x_{0}, \dots, x_{n}] = f [x_{σ (0)}, \dots, x_{σ (n)}]$ where ⁠ $σ$ ⁠ is a permutation that reorders ⁠ $x_{0}, \dots, x_{n}$ ⁠.

Consider the case where ⁠ $σ$ ⁠ reorders ⁠ $x_{0}, \dots, x_{n}$ ⁠ such that identical nodes become contiguous: $(x_{σ (0)}, \dots, x_{σ (n)}) = (\underset{μ_{1} times}{\underset{⏟}{z_{1}, \dots, z_{1}}}, \underset{μ_{2} times}{\underset{⏟}{z_{2}, \dots, z_{2}}}, \dots, \underset{μ_{m} times}{\underset{⏟}{z_{m}, \dots, z_{m}}}) .$ In this case, the divided difference table is constructed using the standard recursive formula whenever the outer nodes are distinct. When evaluating divided differences on ⁠ $i + 1$ ⁠ identical nodes ⁠ $z_{j}$ ⁠, the calculation reduces to and coincides with a scaled higher-order derivative of ⁠ $f$ ⁠: $f [\underset{i + 1 times}{\underset{⏟}{z_{j}, \dots, z_{j}}}] = \frac{f^{(i)} (z_{j})}{i!} .$ All other entries in the divided difference table are then computed using the standard recursive rule.

Generalized case example

To evaluate a confluent divided difference with non-adjacent repeated nodes, such as ⁠ $f [z_{1}, z_{2}, z_{1}]$ ⁠, the node sequence is first permuted to group identical nodes contiguously: $f [z_{1}, z_{2}, z_{1}] = f [z_{1}, z_{1}, z_{2}] .$ The evaluation then proceeds by constructing the divided difference table: $\begin{matrix} x_{0} = z_{1} & f (x_{0}) \\ \frac{f^{'} (z_{1})}{1!} \\ x_{1} = z_{1} & f (x_{1}) & \frac{\frac{f (x_{2}) - f (x_{1})}{x_{2} - x_{1}} - \frac{f^{'} (z_{1})}{1!}}{x_{2} - x_{0}} \\ \frac{f (x_{2}) - f (x_{1})}{x_{2} - x_{1}} \\ x_{2} = z_{2} & f (x_{2}) \end{matrix}$ Extracting and simplifying the rightmost entry from the resulting divided difference table yields: $f [z_{1}, z_{2}, z_{1}] = \frac{f (z_{2}) - f (z_{1})}{(z_{2} - z_{1})^{2}} - \frac{f^{'} (z_{1})}{z_{2} - z_{1}} .$

References

↑ Isaacson, Walter (2014). The Innovators. Simon & Schuster. p. 20. ISBN 978-1-4767-0869-0.
↑ de Boor 2005.
↑ Opitz, G. Steigungsmatrizen, Z. Angew. Math. Mech. (1964), 44, T52–T54
↑ Skof, Fulvia (2011-04-30). Giuseppe Peano between Mathematics and Logic: Proceeding of the International Conference in honour of Giuseppe Peano on the 150th anniversary of his birth and the centennial of the Formulario Mathematico Torino (Italy) October 2-3, 2008. Springer Science & Business Media. p. 40. ISBN 978-88-470-1836-5.
↑ de Boor 2005, p. 65.
↑ Higham 2008, Eq. (B.25).
↑ de Boor 2005, Eq. (51).
↑ Higham 2008, Eq. (B.29).
↑ Burden, Richard L.; Faires, J. Douglas (2011). Numerical Analysis (9th ed.). Cengage Learning. p. 129. ISBN 9780538733519.
↑ Dubeau 2020, p. 269.
↑ Abel, Ulrich; Ivan, Mircea; Riedel, Thomas (2004). "The mean value theorem of Flett and divided differences". Journal of Mathematical Analysis and Applications. 295 (1): 3. doi:10.1016/j.jmaa.2003.10.046. ISSN 0022-247X.
↑ de Boor 2005, p. 48.
↑ Dubeau 2020, p. 259.
↑ de Boor 2005, p. 56.
↑ Higham 2008, Appendix B.16.

Louis Melville Milne-Thomson (2000) [1933]. The Calculus of Finite Differences. American Mathematical Soc. Chapter 1: Divided Differences. ISBN 978-0-8218-2107-7.
Myron B. Allen; Eli L. Isaacson (1998). Numerical Analysis for Applied Science. John Wiley & Sons. Appendix A. ISBN 978-1-118-03027-1.
Ron Goldman (2002). Pyramid Algorithms: A Dynamic Programming Approach to Curves and Surfaces for Geometric Modeling. Morgan Kaufmann. Chapter 4:Newton Interpolation and Difference Triangles. ISBN 978-0-08-051547-2.
de Boor, Carl (2005). "Divided Differences". Surveys in Approximation Theory. 1: 46–69. arXiv:math/0502036. ISSN 1555-578X.
Higham, Nicholas J. (2008). Functions of Matrices: Theory and Computation. Philadelphia, PA, USA: Society for Industrial and Applied Mathematics. doi:10.1137/1.9780898717778. ISBN 978-0-89871-646-7.
Dubeau, François (2020). "On Hermite interpolation and divided differences" (PDF). Surveys in Mathematics and its Applications. 15: 257–279. ISSN 1842-6298.

External links

An implementation in Haskell.

[Isaacson2014-1] Isaacson, Walter (2014). The Innovators. Simon & Schuster. p. 20. ISBN 978-1-4767-0869-0.

[FOOTNOTEde_Boor2005-2] Boor 2005.

[3] Opitz, G. Steigungsmatrizen, Z. Angew. Math. Mech. (1964), 44, T52–T54

[4] Skof, Fulvia (2011-04-30). Giuseppe Peano between Mathematics and Logic: Proceeding of the International Conference in honour of Giuseppe Peano on the 150th anniversary of his birth and the centennial of the Formulario Mathematico Torino (Italy) October 2-3, 2008. Springer Science & Business Media. p. 40. ISBN 978-88-470-1836-5.

[FOOTNOTEde_Boor200565-5] Boor 2005, p. 65.

[FOOTNOTEHigham2008Eq._(B.25)-6] Higham 2008, Eq. (B.25).

[FOOTNOTEde_Boor2005Eq._(51)-7] Boor 2005, Eq. (51).

[FOOTNOTEHigham2008Eq._(B.29)-8] Higham 2008, Eq. (B.29).

[9] Burden, Richard L.; Faires, J. Douglas (2011). Numerical Analysis (9th ed.). Cengage Learning. p. 129. ISBN 9780538733519.

[FOOTNOTEDubeau2020269-10] Dubeau 2020, p. 269.

[11] Abel, Ulrich; Ivan, Mircea; Riedel, Thomas (2004). "The mean value theorem of Flett and divided differences". Journal of Mathematical Analysis and Applications. 295 (1): 3. doi:10.1016/j.jmaa.2003.10.046. ISSN 0022-247X.

[FOOTNOTEde_Boor200548-12] Boor 2005, p. 48.

[FOOTNOTEDubeau2020259-13] Dubeau 2020, p. 259.

[FOOTNOTEde_Boor200556-14] Boor 2005, p. 56.

[FOOTNOTEHigham2008Appendix_B.16-15] Higham 2008, Appendix B.16.