# Tikhonov regularization

Template:Regression bar Tikhonov regularization, named for Andrey Tikhonov, is the most commonly used method of regularization of ill-posed problems. In statistics, the method is known as ridge regression, and, with multiple independent discoveries, it is also variously known as the Tikhonov–Miller method, the Phillips–Twomey method, the constrained linear inversion method, and the method of linear regularization. It is related to the Levenberg–Marquardt algorithm for non-linear least-squares problems.

When the following problem is not well posed (either because of non-existence or non-uniqueness of ${\displaystyle x}$)

${\displaystyle A\mathbf {x} =\mathbf {b} ,}$

then the standard approach is known as ordinary least squares, which seeks to minimize the sum of squared residuals that can be compactly written as

${\displaystyle \|A\mathbf {x} -\mathbf {b} \|^{2}}$

where ${\displaystyle \left\|\cdot \right\|}$ is the Euclidean norm. This may be due to the system being overdetermined or underdetermined. In the latter case this is no better than the original problem. In order to give preference to a particular solution with desirable properties, the regularization term is included in this minimization:

${\displaystyle \|A\mathbf {x} -\mathbf {b} \|^{2}+\|\Gamma \mathbf {x} \|^{2}}$

for some suitably chosen Tikhonov matrix, ${\displaystyle \Gamma }$. In many cases, this matrix is chosen as a multiple of the identity matrix (${\displaystyle \Gamma =\alpha I}$), giving preference to solutions with smaller norms. In other cases, lowpass operators (e.g., a difference operator or a weighted Fourier operator) may be used to enforce smoothness if the underlying vector is believed to be mostly continuous. This regularization improves the conditioning of the problem, thus enabling a direct numerical solution. An explicit solution, denoted by ${\displaystyle {\hat {x}}}$, is given by:

${\displaystyle {\hat {x}}=(A^{T}A+\Gamma ^{T}\Gamma )^{-1}A^{T}\mathbf {b} }$

The effect of regularization may be varied via the scale of matrix ${\displaystyle \Gamma }$. For ${\displaystyle \Gamma =0}$ this reduces to the unregularized least squares solution provided that (ATA)−1 exists.

## History

Tikhonov regularization has been invented independently in many different contexts. It became widely known from its application to integral equations from the work of Andrey Tikhonov and David L. Phillips. Some authors use the term Tikhonov–Phillips regularization. The finite-dimensional case was expounded by Arthur E. Hoerl, who took a statistical approach, and by Manus Foster, who interpreted this method as a WienerKolmogorov filter. Following Hoerl, it is known in the statistical literature as ridge regression.

## Generalized Tikhonov regularization

For general multivariate normal distributions for ${\displaystyle x}$ and the data error, one can apply a transformation of the variables to reduce to the case above. Equivalently, one can seek an ${\displaystyle x}$ to minimize

${\displaystyle \|Ax-b\|_{P}^{2}+\|x-x_{0}\|_{Q}^{2}\,}$

where we have used ${\displaystyle \left\|x\right\|_{Q}^{2}}$ to stand for the weighted norm ${\displaystyle x^{T}Qx}$ (compare with the Mahalanobis distance). In the Bayesian interpretation ${\displaystyle P}$ is the inverse covariance matrix of ${\displaystyle b}$, ${\displaystyle x_{0}}$ is the expected value of ${\displaystyle x}$, and ${\displaystyle Q}$ is the inverse covariance matrix of ${\displaystyle x}$. The Tikhonov matrix is then given as a factorization of the matrix ${\displaystyle Q=\Gamma ^{T}\Gamma }$ (e.g. the Cholesky factorization), and is considered a whitening filter.

This generalized problem has an optimal solution ${\displaystyle x^{*}}$ which can be solved explicitly using the formula

${\displaystyle x^{*}=(A^{T}PA+Q)^{-1}(A^{T}Pb+Qx_{0}).\,}$

or equivalently

${\displaystyle x^{*}=x_{0}+(A^{T}PA+Q)^{-1}(A^{T}P(b-Ax_{0})).\,}$

## Regularization in Hilbert space

Typically discrete linear ill-conditioned problems result from discretization of integral equations, and one can formulate a Tikhonov regularization in the original infinite-dimensional context. In the above we can interpret ${\displaystyle A}$ as a compact operator on Hilbert spaces, and ${\displaystyle x}$ and ${\displaystyle b}$ as elements in the domain and range of ${\displaystyle A}$. The operator ${\displaystyle A^{*}A+\Gamma ^{T}\Gamma }$ is then a self-adjoint bounded invertible operator.

## Relation to singular value decomposition and Wiener filter

With ${\displaystyle \Gamma =\alpha I}$, this least squares solution can be analyzed in a special way via the singular value decomposition. Given the singular value decomposition of A

${\displaystyle A=U\Sigma V^{T}\,}$

with singular values ${\displaystyle \sigma _{i}}$, the Tikhonov regularized solution can be expressed as

${\displaystyle {\hat {x}}=VDU^{T}b}$

where ${\displaystyle D}$ has diagonal values

${\displaystyle D_{ii}={\frac {\sigma _{i}}{\sigma _{i}^{2}+\alpha ^{2}}}}$

and is zero elsewhere. This demonstrates the effect of the Tikhonov parameter on the condition number of the regularized problem. For the generalized case a similar representation can be derived using a generalized singular value decomposition.

Finally, it is related to the Wiener filter:

${\displaystyle {\hat {x}}=\sum _{i=1}^{q}f_{i}{\frac {u_{i}^{T}b}{\sigma _{i}}}v_{i}}$

## Determination of the Tikhonov factor

The optimal regularization parameter ${\displaystyle \alpha }$ is usually unknown and often in practical problems is determined by an ad hoc method. A possible approach relies on the Bayesian interpretation described below. Other approaches include the discrepancy principle, cross-validation, L-curve method, restricted maximum likelihood and unbiased predictive risk estimator. Grace Wahba proved that the optimal parameter, in the sense of leave-one-out cross-validation minimizes:

${\displaystyle G={\frac {\operatorname {RSS} }{\tau ^{2}}}={\frac {\left\|X{\hat {\beta }}-y\right\|^{2}}{\left[\operatorname {Tr} \left(I-X(X^{T}X+\alpha ^{2}I)^{-1}X^{T}\right)\right]^{2}}}}$

Using the previous SVD decomposition, we can simplify the above expression:

${\displaystyle \operatorname {RSS} =\left\|y-\sum _{i=1}^{q}(u_{i}'b)u_{i}\right\|^{2}+\left\|\sum _{i=1}^{q}{\frac {\alpha ^{2}}{\sigma _{i}^{2}+\alpha ^{2}}}(u_{i}'b)u_{i}\right\|^{2}}$
${\displaystyle \operatorname {RSS} =\operatorname {RSS} _{0}+\left\|\sum _{i=1}^{q}{\frac {\alpha ^{2}}{\sigma _{i}^{2}+\alpha ^{2}}}(u_{i}'b)u_{i}\right\|^{2}}$

and

${\displaystyle \tau =m-\sum _{i=1}^{q}{\frac {\sigma _{i}^{2}}{\sigma _{i}^{2}+\alpha ^{2}}}=m-q+\sum _{i=1}^{q}{\frac {\alpha ^{2}}{\sigma _{i}^{2}+\alpha ^{2}}}}$

## Relation to probabilistic formulation

The probabilistic formulation of an inverse problem introduces (when all uncertainties are Gaussian) a covariance matrix ${\displaystyle C_{M}}$ representing the a priori uncertainties on the model parameters, and a covariance matrix ${\displaystyle C_{D}}$ representing the uncertainties on the observed parameters (see, for instance, Tarantola, 2005 [1]). In the special case when these two matrices are diagonal and isotropic, ${\displaystyle C_{M}=\sigma _{M}^{2}I}$ and ${\displaystyle C_{D}=\sigma _{D}^{2}I}$, and, in this case, the equations of inverse theory reduce to the equations above, with ${\displaystyle \alpha ={\sigma _{D}}/{\sigma _{M}}}$.

## Bayesian interpretation

Template:See Although at first the choice of the solution to this regularized problem may look artificial, and indeed the matrix ${\displaystyle \Gamma }$ seems rather arbitrary, the process can be justified from a Bayesian point of view. Note that for an ill-posed problem one must necessarily introduce some additional assumptions in order to get a unique solution. Statistically, the prior probability distribution of ${\displaystyle x}$ is sometimes taken to be a multivariate normal distribution. For simplicity here, the following assumptions are made: the means are zero; their components are independent; the components have the same standard deviation ${\displaystyle \sigma _{x}}$. The data are also subject to errors, and the errors in ${\displaystyle b}$ are also assumed to be independent with zero mean and standard deviation ${\displaystyle \sigma _{b}}$. Under these assumptions the Tikhonov-regularized solution is the most probable solution given the data and the a priori distribution of ${\displaystyle x}$, according to Bayes' theorem.[1]

If the assumption of normality is replaced by assumptions of homoskedasticity and uncorrelatedness of errors, and if one still assumes zero mean, then the Gauss–Markov theorem entails that the solution is the minimal unbiased estimator.{{ safesubst:#invoke:Unsubst||date=__DATE__ |\$B= {{#invoke:Category handler|main}}{{#invoke:Category handler|main}}[citation needed] }}

## References

1. {{#invoke:citation/CS1|citation |CitationClass=book }}
• {{#invoke:citation/CS1|citation

|CitationClass=book }}

• {{#invoke:Citation/CS1|citation

|CitationClass=journal }}

• {{#invoke:Citation/CS1|citation

|CitationClass=journal }}. Translated in {{#invoke:Citation/CS1|citation |CitationClass=journal }}

• {{#invoke:citation/CS1|citation

|CitationClass=book }}

• Tikhonov A.N., Goncharsky A.V., Stepanov V.V., Yagola A.G., 1995, Numerical Methods for the Solution of Ill-Posed Problems, Kluwer Academic Publishers.
• Tikhonov A.N., Leonov A.S., Yagola A.G., 1998, Nonlinear Ill-Posed Problems, V. 1, V. 2, Chapman and Hall.
• Hansen, P.C., 1998, Rank-deficient and Discrete ill-posed problems, SIAM
• Hoerl AE, 1962, Application of ridge analysis to regression problems, Chemical Engineering Progress, 1958, 54–59.
• {{#invoke:Citation/CS1|citation

|CitationClass=journal }}

|CitationClass=book }}

• Tarantola A, 2005, Inverse Problem Theory (free PDF version), Society for Industrial and Applied Mathematics, ISBN 0-89871-572-5
• Template:Cite paper
• {{#invoke:Citation/CS1|citation

|CitationClass=journal }}