Attribute and case selection for NN classifier through rough sets and naturally inspired algorithms

Supervised classification is one of the most active research fields in the Artificial Intelligence community. Nearest Neighbor (NN) is one of the simplest and most consistently accurate approaches to supervised classification. The training set preprocessing is essential for obtaining high quality classification results. This paper introduces an attribute and case selection algorithm using a hybrid Rough Set Theory and naturally inspired approach to improve the NN performance. The proposed algorithm deals with mixed and incomplete, as well as imbalanced datasets. Its performance was tested over repository databases, showing high classification accuracy while keeping few cases and attributes.