Exhaustive and Heuristic Approaches for Minimizing Dimensionality and Misclassification Cost

We consider a special type of dimensionality reduction classification problem where the decision-making objective is to minimize misclassification cost and attributes (MMCA). We propose a two-stage solution approach for solving the MMCA problem. Using simulated data sets and different misclassification cost matrices, we test our two-stage exhaustive and simulated annealing procedures. Our results indicate that the simulated annealing heuristic approach provides competitive performance on test datasets and is computationally efficient.