Optimal Solution of Linear Inequalities with Applications to Pattern Recognition

An algorithm for the optimal solution of consistent and inconsistent linear inequalities is presented, where the optimality criterion is the maximization of the number of constraints satisfied. In the terminology of pattern recognition, the algorithm finds a linear decision function which minimizes the number of patterns misclassified. The algorithm is developed as a nonenumerative search procedure based on several new results established in this paper. Bounds on the search are also developed and the method is experimentally evaluated and shown to be computationally superior to other techniques for finding minimum-error solutions.