The rate of convergence of Sinkhorn balancing (Q802704): Difference between revisions

From MaRDI portal
Import240304020342 (talk | contribs)
Set profile property.
Set OpenAlex properties.
 
(One intermediate revision by one other user not shown)
Property / cites work
 
Property / cites work: Applications of an inequality in information theory to matrices / rank
 
Normal rank
Property / cites work
 
Property / cites work: Growth transformations for functions on manifolds / rank
 
Normal rank
Property / cites work
 
Property / cites work: The diagonal equivalence of a nonnegative matrix to a stochastic matrix / rank
 
Normal rank
Property / cites work
 
Property / cites work: Generalized Iterative Scaling for Log-Linear Models / rank
 
Normal rank
Property / cites work
 
Property / cites work: On the scaling of multidimensional matrices / rank
 
Normal rank
Property / cites work
 
Property / cites work: Q3254327 / rank
 
Normal rank
Property / cites work
 
Property / cites work: On nonnegative matrices / rank
 
Normal rank
Property / cites work
 
Property / cites work: On matrices with doubly stochastic pattern / rank
 
Normal rank
Property / cites work
 
Property / cites work: The Distribution of Positive Elements in Doubly-Stochastic Matrices / rank
 
Normal rank
Property / cites work
 
Property / cites work: Q3219581 / rank
 
Normal rank
Property / cites work
 
Property / cites work: Q5652137 / rank
 
Normal rank
Property / cites work
 
Property / cites work: Methods for scaling to doubly stochastic form / rank
 
Normal rank
Property / cites work
 
Property / cites work: The Jacobian of a growth transformation / rank
 
Normal rank
Property / cites work
 
Property / cites work: A Relationship Between Arbitrary Positive Matrices and Doubly Stochastic Matrices / rank
 
Normal rank
Property / cites work
 
Property / cites work: Concerning nonnegative matrices and doubly stochastic matrices / rank
 
Normal rank
Property / cites work
 
Property / cites work: Diagonal Equivalence to Matrices with Prescribed Row and Column Sums / rank
 
Normal rank
Property / cites work
 
Property / cites work: Problems Involving Diagonal Products in Nonnegative Matrices / rank
 
Normal rank
Property / cites work
 
Property / cites work: Estimating Nonnegative Matrices from Marginal Data / rank
 
Normal rank
Property / cites work
 
Property / cites work: Convergence of the Iterative Scaling Procedure for Non-Negative Matrices / rank
 
Normal rank
Property / cites work
 
Property / cites work: On a Least Squares Adjustment of a Sampled Frequency Table When the Expected Marginal Totals are Known / rank
 
Normal rank
Property / cites work
 
Property / cites work: Maximum Entropy for Hypothesis Formulation, Especially for Multidimensional Contingency Tables / rank
 
Normal rank
Property / full work available at URL
 
Property / full work available at URL: https://doi.org/10.1016/0024-3795(91)90157-r / rank
 
Normal rank
Property / OpenAlex ID
 
Property / OpenAlex ID: W1966824417 / rank
 
Normal rank

Latest revision as of 11:07, 30 July 2024

scientific article
Language Label Description Also known as
English
The rate of convergence of Sinkhorn balancing
scientific article

    Statements

    The rate of convergence of Sinkhorn balancing (English)
    0 references
    0 references
    1991
    0 references
    Let x be an \(n\times n\) stochastic matrix with no zero columns. Define a function f into the \(n\times n\) stochastic matrices defined by \(f(x)_{ij}=z_{ij}(\sum_{k}z_{ik})^{-1}\) where each \(z_{ij}=x_{ij}(\sum_{k}x_{kj})^{-1}\), \(i,j=1,...,n\). The reviewer and \textit{P. Knopp} [Pac. J. Math. 21, 343-348 (1967; Zbl 0152.014)] have shown that if \(x\neq 0\) is an \(n\times n\) nonnegative matrix with total support, i.e. every \(x_{ij}>0\) lies on a positive diagonal, and if \(x^ 0_{ij}=x_{ij}(\sum_{k}x_{ik})^{-1}\), \(i,j=1,...,n\), then the iteration \(x^{t+1}=f(x^ t),\) \(t=0,1,2,...\), converges to a doubly stochastic limit D. The author shows that there is a nonnegative number \(\beta <1\) which depends only upon the limit D such that for every \(\epsilon >0\) there is a norm \(\|.\|\) for which \(\| x^{t+1}-D\| \leq (\beta +\epsilon)\| x^ t-D\|\) holds for every sufficiently large value of t. For \(\epsilon\), \(0<\epsilon <\gamma -\beta\) where \(\beta \leq \gamma <1\), this becomes \(\| x^{t+1}-D\| \leq \gamma \| x^ t-D\|\) for sufficiently large t. Thus \(x^ t\to D\) geometrically. If \(x^ 0\) only has support, i.e. at least one positive diagonal, the convergence may not be geometric. Consider \(\left( \begin{matrix} 1\\ \alpha_ 0\end{matrix} \quad \begin{matrix} 0\\ 1-\alpha_ 0\end{matrix} \right),\) where \(0<\alpha_ 0<1\). Then \(x^ t=\left( \begin{matrix} 1\\ \alpha_ 0/(1+2t\alpha_ 0)\end{matrix} \quad \begin{matrix} 0\\ (1-j\alpha_ 0+2t\alpha_ 0)/(1+2t\alpha_ 0)\end{matrix} \right)\) which converges to \(\left( \begin{matrix} 1\\ 0\end{matrix} \begin{matrix} 0\\ 1\end{matrix} \right)\) at a rate proportional to 1/t.
    0 references
    rate of convergence
    0 references
    Sinkhorn balancing
    0 references
    doubly stochastic matrix
    0 references
    iteration
    0 references

    Identifiers