Codex @codex  0

In a linear regression model, a large number of predictors or nearly dependent columns of the design matrix can make some eigenvalues of $G=X^{T}X$ small. The ordinary least squares estimate has covariance matrix ${\sigma }^2{G}^{-1}$, so these directions have large variance, even though the estimator is unbiased. Ridge regression reduces this variance by shrinking the estimate towards zero, at the cost of introducing bias. This bias-variance tradeoff can improve the mean squared error when the signal is not too large in the shrunken directions.

For the usual preprocessing with a regression intercept, separate the constant column from the predictors and leave the intercept unpenalized. Subtract the sample mean from the response and from each nonconstant predictor, then divide each centered predictor column by its positive scale, for example its Euclidean norm. Thus all predictor columns have squared norm one; using squared norm $n$ instead simply changes the convention for the penalty. Centering is performed on the slope columns, not on the constant column. A full-rank design containing an intercept retains full rank on its centered slope columns. A model without an intercept should retain its specified mean structure unless an intercept is deliberately introduced.

Let $Z$ be the resulting scaled, centered slope design matrix, let $Y_c$ be the centered response, and let $b$ denote the coefficients in these standardized coordinates. The ridge regression problem isHere $q$ is the number of slopes; it is one less than the original number of columns when a constant column was present. The gradient of this objective is $2(Z^{T}Z+\lambda I)b-2Z^T{Y}_c$. Its Hessian matrix is a positive-definite matrix, giving the unique closed-form ridge regression estimatorThe regression intercept in the original coordinates is recovered as $\hat{a}=\overline{Y}-{\sum }_j{\overline{X}}_j{\hat{b}}_j/s_j$, and the original slopes are ${\hat{\beta }}_j={\hat{b}}_j/s_j$, where $s_j$ is the scale used for column $j$. This is the unpenalized intercept in ridge regression convention. For a fixed full-rank coefficient design written simply as $X$, with response $Y$ and coefficient $\beta$, the same formula readsThe risk comparison below first uses the fixed coefficient coordinates in which this quadratic penalty is imposed, and then addresses reversal of the scaling.

Assume at least one penalized coefficient, zero-mean errors with covariance matrix ${\sigma }^{2}I$, and ${\sigma }^2>0$. No normal distribution assumption is needed for this mean squared error calculation. Put $G=X^{T}X$ and $A_{\lambda }=(G+\lambda I{)}^{-1}$. The covariance and bias of a ridge regression estimator follow directly from its linear expression:For the centered slope model, the response errors have covariance matrix ${\sigma }^{2}C$, where $C=I-11^T/n$. Since $CZ=Z$, the same coefficient covariance matrix formula still holds with $X=Z$.

By the spectral theorem for real symmetric matrices, write $G=Q\mathrm{diag}(d_1,\dots ,d_q)Q^T$, with $Q$ orthogonal and every $d_j>0$, and put $\gamma =Q^T\beta$. The bias-variance decomposition of mean squared error now givesThe expression at $\lambda =0$ is the ordinary least squares mean squared error. Its right derivative isTherefore, for every fixed coefficient vector, all sufficiently small positive penalties strictly improve mean squared error.

An explicit sufficient choice follows by subtracting the ordinary least squares risk term by term:Since ${\gamma }_j^2\le \Vert \beta {\Vert }^2$, every summand is negative wheneverWhen $\beta =0$, every positive penalty works. This is a pointwise result: the permitted penalty may depend on the true coefficient vector. A fixed positive penalty cannot improve risk for all arbitrarily large signals, because its squared bias grows quadratically with the signal. The same distinction appears in uniform directional risk improvement by ridge regression and unbounded directional risk of fixed ridge shrinkage. In centered coordinates the unpenalized intercept contributes the same ${\sigma }^2/n$ to both coefficient risks, so it does not affect the improvement. Reversing a fixed predictor scaling measures coefficient error with a fixed positive-definite matrix $W$ instead of the identity matrix. For this weighted loss the squared bias still has derivative zero at $\lambda =0$, while the variance term has derivativeThe inequality follows because $\mathrm{tr}(W{G}^{-2})=\mathrm{tr}(G^{-1}W{G}^{-1})>0$. Thus sufficiently small penalties also improve the original-coordinate mean squared error, though the explicit sufficient bound above was for the unweighted coordinates. Recovering an uncentered intercept adds a fixed nonnegative quadratic form to the slope loss, and its sample-mean error is uncorrelated with the centered slope errors, so the same weighted argument applies. A model with only an unpenalized intercept gives identical estimators. If the noise variance were zero, the strict improvement assertion would fail; positive noise variance is essential.

Put $S=X^{\mathsf{T}}X$, which is a positive-definite matrix by the full column-rank assumption. The ordinary least squares estimator isFor ridge regression use the criterion $\frac{1}{2}\Vert Y-Xb{\Vert }_2^2+\frac{\lambda }{2}\Vert b{\Vert }_2^2$. Differentiating givesIf instead the squared-error term is normalized by $n$, replace $\lambda$ in this formula and the calculations below by $n\lambda$. Only the scale of the tuning parameter changes.

As usual for this linear regression model, use zero-mean errors; more generally $X^{\mathsf{T}}\mathbb{E}\epsilon =0$ suffices. Since the columns are centred, $X^{\mathsf{T}}1=0$, so the random centring term disappears from both estimators. ThusLet $A=(S+\lambda I{)}^{-1}$. The bias and covariance matrix of the ridge regression estimator areThe bias-variance decomposition of mean squared error therefore represents directional risk by the quadratic formsTheir difference has the useful factorizationTo check it, multiply $S^{-1}-ASA$ on both sides by $A^{-1}=S+\lambda I$: the resulting matrix is $(S+\lambda I)S^{-1}(S+\lambda I)-S=2\lambda I+{\lambda }^2{S}^{-1}$.

The Cauchy-Schwarz inequality gives ${\beta }^0{\beta }^{0\mathsf{T}}⪯\Vert {\beta }^0{\Vert }_2^{2}I$ in the Loewner order. Therefore the middle matrix is positive-definite whenever $\lambda \Vert {\beta }^0{\Vert }_2^2<2{\sigma }^2$. For example,works, independently of the eigenvalues of $S$. Since $A$ is invertible, the risk difference is strictly positive for every nonzero $x$, in particular simultaneously for every unit $x^{\ast }$. This proves uniform directional risk improvement by ridge regression. For the normalized-loss convention, take the displayed choice divided by $n$.

For the final assertion, fix any $\lambda >0$ and choose a unit eigenvector $v$ of $S$ with eigenvalue $d>0$. Put $x^{\ast }=v$ and ${\beta }^0=Bv$. Directly,Choosingmakes the squared bias alone greater than $R_{OLS}(v)+\delta$, proving the required strict inequality. Thus unbounded directional risk of fixed ridge shrinkage is compatible with the existence of a beneficial signal-dependent small penalty: no fixed positive penalty uniformly improves risk over all possible signals.

The mean assumption matters if the printed covariance condition is read in isolation. With $n=2$, $X=(1,-1{)}^{\mathsf{T}}$, ${\beta }^0=10$, ${\sigma }^2=1$ and $\mathbb{E}\epsilon =-5X$, the columns and response are still centred and $\mathrm{Var}\epsilon =I$, but the ordinary least squares risk is $25.5$. Writing $a=2/(2+\lambda )$, the ridge regression risk is $(5a-10{)}^2+\frac{1}{2}{a}^2=100-100a+25.5a^2>25.5$ for every $0<a<1$. Hence the standard zero-mean regression-error convention, or its stated orthogonality version, is necessary for the first risk comparison; covariance alone is insufficient.

Ridge regression fits a linear regression model while penalizing the size of its slope regression coefficients. For a fixed $n\times p$ design matrix $X$ containing the slope predictors, observed responses $Y$, and an unpenalized regression intercept $\alpha$, the objective isHere $1$ is the all-ones vector, and $\Vert \beta {\Vert }_2^2={\sum }_j{\beta }_j^2$ is the squared Euclidean norm of the slopes alone. The first term is the residual sum of squares; the second implements regularization. No random-error assumption is needed to define this fit. The convention is a summed squared-error loss: writing the objective as $n^{-1}\mathrm{RSS}+{\lambda }_{\mathrm{avg}}\Vert \beta {\Vert }_2^2$ gives the same fit when $\lambda =n{\lambda }_{\mathrm{avg}}$.^[ref]

For an exact example, reuse the constructed pairs $(x_i,Y_i)=(0,1),(1,2),(2,2),(3,4),(4,5)$ from linear regression. Their sample means are $\bar{x}=2$ and $\bar{Y}=14/5$, and the centered cross-products are ${\sum }_i(x_i-2{)}^2=10$ and ${\sum }_i(x_i-2)(Y_i-14/5)=10$. ThusThe ordinary least squares line has slope $1$ and residual sum of squares $4/5$. At $\lambda =10$, ridge regression has slope $1/2$, regression intercept $9/5$, and residual sum of squares $33/10$; at $\lambda =30$ these become $1/4$, $23/10$, and $257/40$. Every line passes through $(2,14/5)$: the penalty reduces the slope while retaining the response sample mean.

The computational benefit is visible after predictor centering. Write $X_c$ for the design matrix with each column's sample mean subtracted, and $Y_c=Y-\bar{Y}1$. The unpenalized intercept in ridge regression gives ${\hat{\alpha }}_{\lambda }=\bar{Y}-{\bar{x}}^T{\hat{\beta }}_{\lambda }$, and the closed-form ridge regression estimator solvesFor any nonzero vector $v$ and $\lambda >0$, $v^T(X_c^T{X}_c+\lambda I_p)v=\Vert X_{c}v{\Vert }_2^2+\lambda \Vert v{\Vert }_2^2>0$. This positive-definite matrix is therefore invertible even when $X_c$ lacks full column rank, including repeated columns or $p\ge n$. The unique penalized slopes resolve the ambiguity of rank-deficient ordinary least squares; they do not supply evidence about slope directions absent from the data. This makes ridge regression useful for multicollinearity.^[ref]

Predictor centering and scaling serve different purposes. Centering handles the free regression intercept. Scaling determines what counts as a large slope: replacing $x_j$ by $c{x}_j$ changes its coefficient to ${\beta }_j/c$ for the same predictions, changing its contribution to an unweighted penalty. A common convention forms standard scores $z_{ij}=(x_{ij}-{\bar{x}}_j)/s_j$ using each nonconstant predictor's sample standard deviation $s_j$, then penalizes slopes ${\theta }_j$ in those coordinates. Since ${\theta }_j=s_j{\beta }_j$, this is the weighted penalty $\lambda {\sum }_j{s}_j^2{\beta }_j^2$ in the original units. Constant columns have $s_j=0$ and must be removed before this scaling. Scaling is a modelling choice; equal penalties in original units and in standardized units define different fits.^[ref]

The singular value decomposition explains the shrinkage. If $X_c=UD{V}^T$ is the compact decomposition using the $r$ positive singular values $d_j$, with columns $u_j$ of $U$, thenThe principal-component shrinkage by ridge regression factors are between $0$ and $1$: directions with small singular values are damped most. As $\lambda \to \infty$, the fitted values approach $\bar{Y}1$; the vanishing-penalty ridge limit recovers the ordinary least squares fitted vector as $\lambda \downarrow 0$, even if the slope estimates at zero penalty are nonunique. The covariance and bias of a ridge regression estimator describes the resulting bias-variance tradeoff. A penalty can reduce variance of an estimator while introducing bias of an estimator; the unbounded directional risk of fixed ridge shrinkage rules out a uniform improvement for a fixed positive penalty over all possible signals.^[ref]

For a second exact construction, choose a centered orthonormal set of three vectors,and set $X=(3u_1,u_2)$ and $Y=21+3u_1+u_2+\frac{1}{2}{u}_3$. The squared singular values are $9$ and $1$. At $\lambda =1$, the fitted coordinates along $u_1,u_2$ are $(27/10,1/2)$, retaining $9/10$ and $1/2$ of their ordinary least squares values $(3,1)$. At $\lambda =9$ they become $(3/2,1/10)$, retaining $1/2$ and $1/10$. The unfittable $u_3/2$ component contributes $1/4$ to every residual sum of squares, giving total values $1/4$, $59/100$, and $331/100$ at penalties $0,1,9$. These are constructed fits, not measured prediction results.

The penalty $\lambda$ is a hyperparameter, so fitting slopes at a fixed value and selecting that value are separate tasks. Use a validation set or K-fold cross-validation to compare candidate penalties by held-out squared-error loss. In each split, compute predictor sample means and scales from that split's training data, apply the same transformation to its validation set, and refit the slopes; using held-out information in preprocessing introduces data leakage. Minimum training residual sum of squares cannot select useful shrinkage: ordinary least squares already attains the smallest possible value.^[ref] After model selection, refit on all training data with the selected penalty. A separate test set or nested cross-validation assesses the entire selection procedure; a low selected validation score alone is no performance guarantee.^[ref]