DMI Technical Reports: Recent submissions
Now showing items 2140 of 46

Semismooth Support Vector Machines
(20001129)The linear support vector machine can be posed as a quadratic pro gram in a variety of ways. In this paper, we look at a formulation using the twonorm for the misclassi cation error that leads to a positive de  nite ... 
PrivacyPreserving Linear and Nonlinear Approximation via Linear Programming
(2011)We propose a novel privacypreserving random kernel approximation based on a data matrix A ? Rm�n whose rows are divided into privately owned blocks. Each block of rows belongs to a different entity that is unwilling to ... 
Absolute Value Equation Solution via Dual Complementarity
(2011)By utilizing a dual complementarity condition, we propose an iterative method for solving the NPhard absolute value equation (AVE): Ax?x = b, where A is an n�n square matrix. The algorithm makes no assumptions on the ... 
Equivalence of Minimal L0 and Lp Norm Solutions of Linear Equalities, Inequalities and Linear Programs for Sufficiently Small p
(2011)For a bounded system of linear equalities and inequalities we show that the NPhard ?0 norm minimization problem min x0 subject to Ax = a, Bx ? b and x? ? 1, is completely equivalent to the concave minimization ... 
PrimalDual Bilinear Programming Solution of the Absolute Value Equation
(2011)We propose a finitely terminating primaldual bilinear programming algorithm for the solution of the NPhard absolute value equation (AVE): Ax ? x = b, where A is an n � n square matrix. The algorithm, which makes no ... 
PrivacyPreserving Horizontally Partitioned Linear Programs
(2010)We propose a simple privacypreserving reformulation of a linear program whose equality constraint matrix is partitioned into groups of rows. Each group of matrix rows and its corresponding right hand side vector are ... 
Probability of Unique Integer Solution to a System of Linear Equations
(2009)We consider a system of m linear equations in n variables Ax = d and give necessary and sufficient conditions for the existence of a unique solution to the system that is integer: x ? {?1,1}n. We achieve this by reformulating ... 
PrivacyPreserving Random Kernel Classification of Checkerboard Partitioned Data
(2008)We propose a privacypreserving support vector machine (SVM) classifier for a data matrix A whose input feature columns as well as individual data point rows are divided into groups belonging to different entities. Each ... 
PrivacyPreserving Classification of Horizontally Partitioned Data via Random Kernels
(2007)We propose a novel privacypreserving nonlinear support vector machine (SVM) classifier for a data matrix A whose columns represent input space features and whose individual rows are divided into groups of rows. Each ... 
PrivacyPreserving Classification of Vertically Partitioned Data via Random Kernels
(2007)We propose a novel privacypreserving support vector machine (SVM) classifier for a data matrix A whose input feature columns are divided into groups belonging to different entities. Each entity is unwilling to share its ... 
Exactness Conditions for a Convex Differentiable Exterior Penalty for Linear Programming
(2007)Sufficient conditions are given for a classical dual exterior penalty function of a linear program to be independent of its penalty parameter. This ensures that an exact solution to the primal linear program can be obtained ... 
Chunking for Massive Nonlinear Kernel Classification
(2006)A chunking procedure [2] utilized in [18] for linear classifiers is proposed here for nonlinear kernel classification of massive datasets. A highly accurate algorithm based on nonlinear support vector machines that ... 
Proximal KnowledgeBased Classification
(20080626)Prior knowledge over general nonlinear sets is incor porated into proximal nonlinear kernel classification problems as linear equalities. The key tool in this incorporation is the conversion of general nonlinear prior ... 
Nonlinear KnowledgeBased Classification
(2006)Prior knowledge over general nonlinear sets is incorporated into nonlinear kernel classification problems as linear constraints in a linear program. The key tool in this incorporation is a theorem of the alternative for ... 
Massive Data Classification via Unconstrained Support Vector Machines
(2006)A highly accurate algorithm, based on support vector machines formulated as linear programs [13, 1], is proposed here as a completely unconstrained minimization problem [15]. Combined with a chunking procedure [2] this ... 
Absolute Value Equations
(2005)We investigate existence and nonexistence of solutions for NPhard equations in volving absolute values of variables: Ax ? x = b, where A is an arbitrary n � n real matrix. By utilizing an equivalence relation to the ... 
RSVM: Reduced Support Vector Machines
(200101)An algorithm is proposed which generates a nonlinear kernelbased separating surface that requires as little as 1% of a large dataset for its explicit evaluation. To generate this nonlinear surface, the entire dataset ... 
Lagrangian Support Vector Machines
(2000)An implicit Lagrangian for the dual of a simple reformulation of the standard quadratic program of a linear support vector machine is proposed. This leads to the minimization of an unconstrained di erentiable convex ... 
Interior Point Methods for Massive Support Vector Machines
(20000525)We investigate the use of interior point methods for solving quadratic programming problems with a small number of linear constraints where the quadratic term consists of a lowrank update to a positive semide nite matrix. ... 
A Practical Approach to Samplepath Simulation Optimization
(2000)We propose solving continuous parametric simulation optimizations using a deterministic nonlinear optimiza tion algorithm and samplepath simulations. The op timization problem is written in a modeling language with ...