Linear goal programming in estimation of classification probability

Article Abstract:

Linear programming applications to group membership probabilities do not provide for within group discrimination capability that hinders the use in statistical discriminant analysis. A continuous goal programming (GP) model using conditional properties of group membership has been developed that provides for this feature. The GP uses order preservation of membership probability pairs in the development sample and has been shown to be effective.

---

New LP based heuristics for the classification problem

Article Abstract:

Three linear programming heuristic procedures for use in two group classification problem applications are evaluated. The procedures utilize expected total cost of misclassification minimization and are shown to be more computationally efficient than existing methodologies. In addition, two of the techniques, single seed and three seed heuristics, provide almost optimal solutions over a range of experimental conditions.

Heuristic programming

---

A modified simplex approach for solving bilevel linear programming problems

Article Abstract:

A technique for simultaneously solving linear programming problems of two levels is presented. The method simplifies the problem by replacing the inner level with its necessary and sufficient Kunh-Tucker form. The complementary slackness equations are then considered penalty funtions of the outer level and the resulting quadratic programme can be solved by iterative use of a modified simplex algorithm.

Usage, Simplex method, Simplexes (Mathematics)

Similar abstracts:

• Abstracts: Least absolute deviation estimation of stationary time series models. Multicriteria optimization of a single-cell oil production
• Abstracts: A heuristic for finding embedded network structure in mathematical programmes. The concave cost supply problem

This website is not affiliated with document authors or copyright owners. This page is provided for informational purposes only. Unintentional errors are possible.