The noisy multidimensional scaling problem: an optimization approach
نویسنده
چکیده
Multidimensional scaling is a fundamental problem in data analysis and have a lot of applications. It’s goal is to look for an Euclidean graphic representation of a given set of data in a “low’ dimensional space (generally in IR or IR). This problem can be formulated as a nonlinear global optimization problem. To solve it, a Lenvenberg-Marquardt method is used upon different cost functions. Results are given on some classical test examples. The limits of such a local method carry us to consider a deterministic global algorithm using Interval techniques. Key-Words; Multidimensional Scaling (MDS), Data Analysis, Levenberg-Marquardt, Interval Arithmetic, Global Optimization
منابع مشابه
Some Heuristic and Deterministic Global Optimization Algorithms to Solve the Problem of Multidimensional Scaling
Multidimensional scaling (MDS) which is a fundamental problem in data analysis, can be formulated as a nonlinear global optimization problem. In order to solve it, some local, heuristic and global optimization methods are implemented. In another hand, to perform our new heuristic approach, the MDS problem is studied by using a multistart Levenberg-Marquardt method and, also for some small examp...
متن کاملAn improved genetic algorithm for multidimensional optimization of precedence-constrained production planning and scheduling
Integration of production planning and scheduling is a class of problems commonly found in manufacturing industry. This class of problems associated with precedence constraint has been previously modeled and optimized by the authors, in which, it requires a multidimensional optimization at the same time: what to make, how many to make, where to make and the order to make. It is a combinatorial,...
متن کاملQuadratic Programming with Complementarity Constraints for Multidimensional Scaling with City-Block Distances
In this paper, we consider an optimization problem arising in multidimensional scaling with city-block distances. The objective function of this problem has many local minimum points and may be even non-differentiable at a minimum point. We reformulate the problem into a problem with convex quadratic objective function, linear and complementarity constraints. In addition, we propose an algorith...
متن کاملKernel Regularization and Dimension Reduction
It is often possible to use expert knowledge or other sources of information to obtain dissimilarity measures for pairs of objects, which serve as pseudo-distances between the objects. When dissimilarity information is available as the data, there are two different types of problems of interest. The first is to estimate full position configuration for all objects in a low dimensional space whil...
متن کاملExtensions of Classical Multidimensional Scaling: Variable Alternation and Nonconvex Duality∗
Classical multidimensional scaling constructs a configuration of points that minimizes a certain measure of discrepancy between the configuration’s interpoint distance matrix and a fixed dissimilarity matrix. Recent extensions have replaced the fixed dissimilarity matrix with a closed and convex set of dissimilarity matrices. These extensions lead to optimization problems with two sets of decis...
متن کامل