Editing Interpolation

{{Short description|Method for estimating new data within known data points}}
{{Other uses}}
{{distinguish|Interpellation (disambiguation){{!}}Interpellation}}

{{more footnotes|date=October 2016}}
In the [[mathematics|mathematical]] field of [[numerical analysis]], '''interpolation''' is a type of [[estimation]], a method of constructing (finding) new [[data points]] based on the range of a [[discrete set]] of known data points.<ref>{{cite EB1911 |wstitle=Interpolation |volume=14 |pages=706–710 |first=William Fleetwood |last=Sheppard |author-link=William Fleetwood Sheppard}}</ref><ref>{{Cite book|last=Steffensen|first=J. F.|url=https://www.worldcat.org/oclc/867770894|title=Interpolation|date=2006|isbn=978-0-486-15483-1|edition=Second|location=Mineola, N.Y.|oclc=867770894}}</ref>

In [[engineering]] and [[science]], one often has a number of data points, obtained by [[sampling (statistics)|sampling]] or [[experimentation]], which represent the values of a function for a limited number of values of the [[Dependent and independent variables|independent variable]]. It is often required to '''interpolate'''; that is, estimate the value of that function for an intermediate value of the independent variable.

A closely related problem is the [[function approximation|approximation]] of a complicated function by a simple function. Suppose the formula for some given function is known, but too complicated to evaluate efficiently. A few data points from the original function can be interpolated to produce a simpler function which is still fairly close to the original. The resulting gain in simplicity may outweigh the loss from interpolation error and give better performance in calculation process.

[[File:Splined epitrochoid.svg|300px|thumb|An interpolation of a finite set of points on an [[epitrochoid]]. The points in red are connected by blue interpolated [[spline (mathematics)|spline curves]] deduced only from the red points. The interpolated curves have polynomial formulas much simpler than that of the original epitrochoid curve.]]

==Example==
This table gives some values of an unknown function <math>f(x)</math>.
[[File:Interpolation Data.svg|right|thumb|230px|Plot of the data points as given in the table]]
{| cellpadding=0 cellspacing=0
|width="20px"|
! <math>x</math>
|width="10px"|
!colspan=3 align=center| <math>f(x)</math>
|-
| || 0 || ||align=right| 0
|-
| || 1 || ||align=right| 0 || . || 8415
|-
| || 2 || ||align=right| 0 || . || 9093
|-
| || 3 || ||align=right| 0 || . ||1411
|-
| || 4 || ||align=right| &minus;0 || . || 7568
|-
| || 5 || ||align=right| &minus;0 || . || 9589
|-
| || 6 || ||align=right| &minus;0 || . || 2794
|}
Interpolation provides a means of estimating the function at intermediate points, such as <math>x=2.5.</math>

We describe some [[algorithm|methods]] of interpolation, differing in such properties as: accuracy, cost, number of data points needed, and [[smooth function|smoothness]] of the resulting [[interpolant]] function.
{{clear}}

===Piecewise constant interpolation===
[[File:Piecewise constant.svg|thumb|right|Piecewise constant interpolation, or [[nearest-neighbor interpolation]]]]
{{Further|Nearest-neighbor interpolation}}
The simplest interpolation method is to locate the nearest data value, and assign the same value. In simple problems, this method is unlikely to be used, as [[linear]] interpolation (see below) is almost as easy, but in higher-dimensional [[multivariate interpolation]], this could be a favourable choice for its speed and simplicity.
{{clear}}

===Linear interpolation===
[[File:Interpolation example linear.svg|right|thumb|230px|Plot of the data with linear interpolation superimposed]]
{{Main|Linear interpolation}}
One of the simplest methods is linear interpolation (sometimes known as lerp). Consider the above example of estimating ''f''(2.5). Since 2.5 is midway between 2 and 3, it is reasonable to take ''f''(2.5) midway between ''f''(2) = 0.9093 and ''f''(3) = 0.1411, which yields 0.5252.

Generally, linear interpolation takes two data points, say (''x''<sub>''a''</sub>,''y''<sub>''a''</sub>) and (''x''<sub>''b''</sub>,''y''<sub>''b''</sub>), and the interpolant is given by:
:<math> y = y_a + \left( y_b-y_a \right) \frac{x-x_a}{x_b-x_a} \text{ at the point } \left( x,y \right) </math>

:<math> \frac{y-y_a}{y_b-y_a} = \frac{x-x_a}{x_b-x_a} </math>

:<math> \frac{y-y_a}{x-x_a} = \frac{y_b-y_a}{x_b-x_a} </math>

This previous equation states that the slope of the new line between <math> (x_a,y_a) </math> and <math> (x,y) </math> is the same as the slope of the line between <math> (x_a,y_a) </math> and <math> (x_b,y_b) </math>

Linear interpolation is quick and easy, but it is not very precise. Another disadvantage is that the interpolant is not [[derivative|differentiable]] at the point ''x''<sub>''k''</sub>.

The following error estimate shows that linear interpolation is not very precise. Denote the function which we want to interpolate by ''g'', and suppose that ''x'' lies between ''x''<sub>''a''</sub> and ''x''<sub>''b''</sub> and that ''g'' is twice continuously differentiable. Then the linear interpolation error is

:<math> |f(x)-g(x)| \le C(x_b-x_a)^2 \quad\text{where}\quad C = \frac18 \max_{r\in[x_a,x_b]} |g''(r)|. </math>

In words, the error is proportional to the square of the distance between the data points. The error in some other methods, including [[polynomial interpolation]] and spline interpolation (described below), is proportional to higher powers of the distance between the data points. These methods also produce smoother interpolants.
{{clear}}

===Polynomial interpolation===
[[File:Interpolation example polynomial.svg|right|thumb|230px|Plot of the data with polynomial interpolation applied]]
{{Main|Polynomial interpolation}}
Polynomial interpolation is a generalization of linear interpolation. Note that the linear interpolant is a [[linear function]]. We now replace this interpolant with a [[polynomial]] of higher [[degree of a polynomial|degree]].

Consider again the problem given above. The following sixth degree polynomial goes through all the seven points:
:<math> f(x) = -0.0001521 x^6 - 0.003130 x^5 + 0.07321 x^4 - 0.3577 x^3 + 0.2255 x^2 + 0.9038 x. </math>
<!-- Coefficients are 0, 0.903803333333334, 0.22549749999997, -0.35772291666664, 0.07321458333332, -0.00313041666667, -0.00015208333333. -->
Substituting ''x'' = 2.5, we find that ''f''(2.5) = ~0.59678.

Generally, if we have ''n'' data points, there is exactly one polynomial of degree at most ''n''&minus;1 going through all the data points. The interpolation error is proportional to the distance between the data points to the power ''n''. Furthermore, the interpolant is a polynomial and thus infinitely differentiable. So, we see that polynomial interpolation overcomes most of the problems of linear interpolation.

However, polynomial interpolation also has some disadvantages. Calculating the interpolating polynomial is computationally expensive (see [[Computational complexity theory|computational complexity]]) compared to linear interpolation. Furthermore, polynomial interpolation may exhibit oscillatory artifacts, especially at the end points (see [[Runge's phenomenon]]).

Polynomial interpolation can estimate local maxima and minima that are outside the range of the samples, unlike linear interpolation. For example, the interpolant above has a local maximum at ''x'' ≈ 1.566, ''f''(''x'') ≈ 1.003 and a local minimum at ''x'' ≈ 4.708, ''f''(''x'') ≈ −1.003. However, these maxima and minima may exceed the theoretical range of the function; for example, a function that is always positive may have an interpolant with negative values, and whose inverse therefore contains false [[division by zero|vertical asymptotes]].

More generally, the shape of the resulting curve, especially for very high or low values of the independent variable, may be contrary to commonsense; that is, to what is known about the experimental system which has generated the data points. These disadvantages can be reduced by using spline interpolation or restricting attention to [[Chebyshev polynomials]].
{{clear}}

===Spline interpolation===
[[File:Interpolation example spline.svg|right|thumb|230px|Plot of the data with spline interpolation applied]]
{{Main|Spline interpolation}}

Linear interpolation uses a linear function for each of intervals [''x''<sub>''k''</sub>,''x''<sub>''k+1''</sub>]. Spline interpolation uses low-degree polynomials in each of the intervals, and chooses the polynomial pieces such that they fit smoothly together. The resulting function is called a spline.

For instance, the [[natural cubic spline]] is [[piecewise]] cubic and twice continuously differentiable. Furthermore, its second derivative is zero at the end points. The natural cubic spline interpolating the points in the table above is given by

: <math> f(x) = \begin{cases}
-0.1522 x^3 + 0.9937 x, & \text{if } x \in [0,1], \\
-0.01258 x^3 - 0.4189 x^2 + 1.4126 x - 0.1396, & \text{if } x \in [1,2], \\
0.1403 x^3 - 1.3359 x^2 + 3.2467 x - 1.3623, & \text{if } x \in [2,3], \\
0.1579 x^3 - 1.4945 x^2 + 3.7225 x - 1.8381, & \text{if } x \in [3,4], \\
0.05375 x^3 -0.2450 x^2 - 1.2756 x + 4.8259, & \text{if } x \in [4,5], \\
-0.1871 x^3 + 3.3673 x^2 - 19.3370 x + 34.9282, & \text{if } x \in [5,6].
\end{cases} </math>

In this case we get ''f''(2.5)&nbsp;=&nbsp;0.5972.

Like polynomial interpolation, spline interpolation incurs a smaller error than linear interpolation, while the interpolant is smoother and easier to evaluate than the high-degree polynomials used in polynomial interpolation. However, the global nature of the basis functions leads to ill-conditioning. This is completely mitigated by using splines of compact support, such as are implemented in Boost.Math and discussed in Kress.<ref>{{cite book|last1=Kress|first1=Rainer|title=Numerical Analysis|url=https://archive.org/details/springer_10.1007-978-1-4612-0599-9|year=1998|publisher=Springer |isbn=9781461205999}}</ref>
{{Clear}}

=== Mimetic interpolation ===
{{Main|Mimetic interpolation}}
Depending on the underlying discretisation of fields, different interpolants may be required. In contrast to other interpolation methods, which estimate functions on target points, mimetic interpolation evaluates the integral of fields on target lines, areas or volumes, depending on the type of field (scalar, vector, pseudo-vector or pseudo-scalar).

A key feature of mimetic interpolation is that [[vector calculus identities]] are satisfied, including [[Stokes' theorem]] and the [[divergence theorem]]. As a result, mimetic interpolation conserves line, area and volume integrals.<ref>{{Cite journal |last1=Pletzer |first1=Alexander |last2=Hayek |first2=Wolfgang |date=2019-01-01 |title=Mimetic Interpolation of Vector Fields on Arakawa C/D Grids |url=https://journals.ametsoc.org/view/journals/mwre/147/1/mwr-d-18-0146.1.xml |journal=Monthly Weather Review |language=EN |volume=147 |issue=1 |pages=3–16 |doi=10.1175/MWR-D-18-0146.1 |bibcode=2019MWRv..147....3P |s2cid=125214770 |issn=1520-0493 |access-date=2022-06-07 |archive-date=2022-06-07 |archive-url=https://web.archive.org/web/20220607041035/https://journals.ametsoc.org/view/journals/mwre/147/1/mwr-d-18-0146.1.xml |url-status=live }}</ref> Conservation of line integrals might be desirable when interpolating the [[electric field]], for instance, since the line integral gives the [[electric potential]] difference at the endpoints of the integration path.<ref>{{Citation |last1=Stern |first1=Ari |title=Geometric Computational Electrodynamics with Variational Integrators and Discrete Differential Forms |date=2015 |url=http://link.springer.com/10.1007/978-1-4939-2441-7_19 |work=Geometry, Mechanics, and Dynamics |volume=73 |pages=437–475 |editor-last=Chang |editor-first=Dong Eui |place=New York, NY |publisher=Springer New York |doi=10.1007/978-1-4939-2441-7_19 |isbn=978-1-4939-2440-0 |access-date=2022-06-15 |last2=Tong |first2=Yiying |last3=Desbrun |first3=Mathieu |last4=Marsden |first4=Jerrold E. |series=Fields Institute Communications |s2cid=15194760 |editor2-last=Holm |editor2-first=Darryl D. |editor3-last=Patrick |editor3-first=George |editor4-last=Ratiu |editor4-first=Tudor|arxiv=0707.4470 }}</ref> Mimetic interpolation ensures that the error of estimating the line integral of an electric field is the same as the error obtained by interpolating the potential at the end points of the integration path, regardless of the length of the integration path. 

[[Linear interpolation|Linear]], [[Bilinear interpolation|bilinear]] and [[trilinear interpolation]] are also considered mimetic, even if it is the field values that are conserved (not the integral of the field). Apart from linear interpolation, area weighted interpolation can be considered one of the first mimetic interpolation methods to have been developed.<ref>{{Cite journal |last=Jones |first=Philip |title=First- and Second-Order Conservative Remapping Schemes for Grids in Spherical Coordinates |journal=Monthly Weather Review |year=1999 |volume=127 |issue=9 |pages=2204–2210|doi=10.1175/1520-0493(1999)127<2204:FASOCR>2.0.CO;2 |bibcode=1999MWRv..127.2204J |s2cid=122744293 |doi-access=free }}</ref>

==Functional interpolation==
The [[Theory of functional connections|Theory of Functional Connections]] (TFC) is a mathematical framework specifically developed for [https://www.mdpi.com/journal/mathematics/sections/functional_interpolation functional interpolation]. Given any interpolant that satisfies a set of constraints, TFC derives a functional that represents the entire family of interpolants satisfying those constraints, including those that are discontinuous or partially defined. These functionals identify the subspace of functions where the solution to a constrained optimization problem resides. Consequently, TFC transforms constrained optimization problems into equivalent unconstrained formulations. This transformation has proven highly effective in the solution of [[Differential equation|differential equations]]. TFC achieves this by constructing a constrained functional (a function of a free function), that inherently satisfies given constraints regardless of the expression of the free function. This simplifies solving various types of equations and significantly improves the efficiency and accuracy of methods like [[Physics-informed neural networks|Physics-Informed Neural Networks]] (PINNs). TFC offers advantages over traditional methods like [[Lagrange multiplier|Lagrange multipliers]] and [[spectral method]]s by directly addressing constraints analytically and avoiding iterative procedures, although it cannot currently handle inequality constraints.

==Function approximation==
Interpolation is a common way to approximate functions. Given a function <math>f:[a,b] \to \mathbb{R}</math> with a set of points <math>x_1, x_2, \dots, x_n \in [a, b]</math> one can form a function <math>s: [a,b] \to \mathbb{R}</math> such that <math>f(x_i)=s(x_i)</math> for <math>i=1, 2, \dots, n</math> (that is, that <math>s</math> interpolates <math>f</math> at these points). In general, an interpolant need not be a good approximation, but there are well known and often reasonable conditions where it will. For example, if <math>f\in C^4([a,b])</math> (four times continuously differentiable) then [[spline interpolation|cubic spline interpolation]] has an error bound given by <math>\|f-s\|_\infty \leq C \|f^{(4)}\|_\infty h^4</math> where <math>h \max_{i=1,2, \dots, n-1} |x_{i+1}-x_i|</math> and <math>C</math> is a constant.<ref>{{cite journal |last1=Hall |first1=Charles A. |last2=Meyer |first2=Weston W. |title=Optimal Error Bounds for Cubic Spline Interpolation |journal=Journal of Approximation Theory |date=1976 |volume=16 |issue=2 |pages=105–122 |doi=10.1016/0021-9045(76)90040-X |doi-access=free }}</ref>

==Via Gaussian processes==
[[Gaussian process]] is a powerful non-linear interpolation tool. Many popular interpolation tools are actually equivalent to particular Gaussian processes. Gaussian processes can be used not only for fitting an interpolant that passes exactly through the given data points but also for regression; that is, for fitting a curve through noisy data. In the geostatistics community Gaussian process regression is also known as [[Kriging]].

==Inverse Distance Weighting==
[[Inverse distance weighting|Inverse Distance Weighting]] (IDW) is a spatial interpolation method that estimates values based on nearby data points, with closer points having more influence.<ref>{{cite journal |last1=Donald |first1=Shepard |title=A two-dimensional interpolation function for irregularly-spaced data |journal=23rd ACM National Conference |date=1968}}</ref> It uses an inverse power law for weighting, where higher power values emphasize local effects, while lower values create a smoother surface. IDW is widely used in [[Geographic information system|GIS]], [[meteorology]], and environmental modeling for its simplicity but may produce artifacts in clustered or uneven data.<ref>{{cite journal |last1=Ben Moshe |first1=Nir |title=A Simple Solution for the Inverse Distance Weighting Interpolation (IDW) Clustering Problem |journal=Sci |date=2025 |volume=7 |issue=1 |page=30 |doi=10.3390/sci7010030|doi-access=free }}</ref>

==Other forms==
Other forms of interpolation can be constructed by picking a different class of interpolants. For instance, rational interpolation is '''interpolation''' by [[rational function]]s using [[Padé approximant]], and [[trigonometric interpolation]] is interpolation by [[trigonometric polynomial]]s using [[Fourier series]]. Another possibility is to use [[wavelet]]s.

The [[Whittaker–Shannon interpolation formula]] can be used if the number of data points is infinite or if the function to be interpolated has compact support.

Sometimes, we know not only the value of the function that we want to interpolate, at some points, but also its derivative. This leads to [[Hermite interpolation]] problems.

When each data point is itself a function, it can be useful to see the interpolation problem as a partial [[advection]] problem between each data point. This idea leads to the [[displacement interpolation]] problem used in [[Transportation theory (mathematics)|transportation theory]].

==In higher dimensions==
{{comparison_of_1D_and_2D_interpolation.svg|250px|}}
{{Main|Multivariate interpolation}}
Multivariate interpolation is the interpolation of functions of more than one variable. 
Methods include [[nearest-neighbor interpolation]], [[bilinear interpolation]] and [[bicubic interpolation]] in two dimensions, and [[trilinear interpolation]] in three dimensions.
They can be applied to gridded or scattered data. Mimetic interpolation generalizes to <math>n</math> dimensional spaces where <math>n > 3</math>.<ref>{{Cite book |last=Whitney |first=Hassler |title=Geometric Integration Theory |publisher=Dover Books on Mathematics |year=1957 |isbn=978-0486445830}}</ref><ref>{{Cite journal |last1=Pletzer |first1=Alexander |last2=Fillmore |first2=David |title=Conservative interpolation of edge and face data on n dimensional structured grids using differential forms |journal=Journal of Computational Physics |year=2015 |volume=302 |pages=21–40 |doi=10.1016/j.jcp.2015.08.029 |bibcode=2015JCoPh.302...21P |doi-access=free }}</ref>

<gallery>
Image:Nearest2DInterpolExample.png|Nearest neighbor
Image:BilinearInterpolExample.png|Bilinear
Image:BicubicInterpolationExample.png|Bicubic
</gallery>

==In digital signal processing==
In the domain of digital signal processing, the term interpolation refers to the process of converting a sampled digital signal (such as a sampled audio signal) to that of a higher sampling rate ([[Upsampling]]) using various digital filtering techniques (for example, convolution with a frequency-limited impulse signal). In this application there is a specific requirement that the harmonic content of the original signal be preserved without creating aliased harmonic content of the original signal above the original [[Nyquist frequency|Nyquist limit]] of the signal (that is, above fs/2 of the original signal sample rate). An early and fairly elementary discussion on this subject can be found in Rabiner and Crochiere's book ''Multirate Digital Signal Processing''.<ref>{{Cite book |title=R.E. Crochiere and L.R. Rabiner. (1983). Multirate Digital Signal Processing. Englewood Cliffs, NJ: Prentice–Hall. |isbn=0136051626 |last1=Crochiere |first1=Ronald E. |last2=Rabiner |first2=Lawrence R. |date=1983 |publisher=Prentice-Hall }}</ref>

==Related concepts==
The term ''[[extrapolation]]'' is used to find data points outside the range of known data points.

In [[curve fitting]] problems, the constraint that the interpolant has to go exactly through the data points is relaxed. It is only required to approach the data points as closely as possible (within some other constraints).  This requires parameterizing the potential interpolants and having some way of measuring the error.  In the simplest case this leads to [[least squares]] approximation.

[[Approximation theory]] studies how to find the best approximation to a given function by another function from some predetermined class, and how good this approximation is. This clearly yields a bound on how well the interpolant can approximate the unknown function.

==Generalization==
If we consider <math>x</math> as a variable in a [[topological space]],  and the function <math>f(x)</math> mapping to a [[Banach space]], then the problem is treated as "interpolation of operators".<ref>Colin Bennett, Robert C. Sharpley, ''Interpolation of Operators'', Academic Press 1988</ref> The classical results about interpolation of operators are the [[Riesz–Thorin theorem]] and the [[Marcinkiewicz theorem]]. There are also many other subsequent results.

==See also==
{{Div col|colwidth=30em}}
* [[Barycentric coordinate system|Barycentric coordinates]] – for interpolating within on a triangle or tetrahedron
* [[Brahmagupta's interpolation formula]]
* [[Discretization]]
* [[Fractal compression#Fractal interpolation|Fractal interpolation]]
* [[Imputation (statistics)]]
* [[Lagrange polynomial|Lagrange interpolation]]
* [[Missing data]]
* [[Newton–Cotes formulas]]
* [[Radial basis function interpolation]]
* [[Simple rational approximation]]
* [[Smoothing]]
{{div col end}}

==References==
{{Reflist}}

==External links==
{{Commons category}}
* Online tools for [http://tools.timodenk.com/linear-interpolation linear] {{Webarchive|url=https://web.archive.org/web/20160918103516/http://tools.timodenk.com/linear-interpolation |date=2016-09-18 }}, [http://tools.timodenk.com/quadratic-interpolation quadratic] {{Webarchive|url=https://web.archive.org/web/20160918102633/http://tools.timodenk.com/quadratic-interpolation |date=2016-09-18 }}, [http://tools.timodenk.com/cubic-spline-interpolation cubic spline] {{Webarchive|url=https://web.archive.org/web/20160820175607/http://tools.timodenk.com/cubic-spline-interpolation |date=2016-08-20 }}, and [http://tools.timodenk.com/polynomial-interpolation polynomial] {{Webarchive|url=https://web.archive.org/web/20160918102129/http://tools.timodenk.com/polynomial-interpolation |date=2016-09-18 }} interpolation with visualisation and [[JavaScript]] source code.
* [http://sol.gfxile.net/interpolation/index.html Sol Tutorials - Interpolation Tricks] {{Webarchive|url=https://web.archive.org/web/20210131063823/http://sol.gfxile.net/interpolation/index.html |date=2021-01-31 }}
* [http://www.boost.org/doc/libs/release/libs/math/doc/html/math_toolkit/barycentric.html Barycentric rational interpolation in Boost.Math]
* [http://www.boost.org/doc/libs/release/libs/math/doc/html/math_toolkit/sf_poly/chebyshev.html Interpolation via the Chebyshev transform in Boost.Math]

{{Authority control}}

[[Category:Interpolation| ]]
[[Category:Video]]
[[Category:Video signal]]