PARTITIONING OF FEATURE SPACE FOR PATTERN-CLASSIFICATION

The article proposes a simple approach for finding a fuzzy partitionin g of a feature space for pattern classification problems. A feature sp ace is initially decomposed into some overlapping hyperboxes depending on the relative positions of the pattern classes found in the trainin g samples. A few fuzzy if-then rules reflecting the pattern classes by the generated hyperboxes are then obtained in terms of a relational m atrix. The relational matrix is utilized in the modified compositional rule of inference in order to recognize an unknown pattern. The propo sed system is capable of handling imprecise information both in the le arning and the processing phases. The imprecise information is conside red to be either incomplete or mixed or interval or linguistic in form . Ways of handling such imprecise information are also discussed. The effectiveness of the system is demonstrated on some synthetic data set s in two-dimensional feature space. The practical applicability of the system is verified on four real data such as the Iris data set, an ap pendicitis data set, a speech data set and a hepatic disease data set. (C) 1997 Pattern Recognition Society. Published by Elsevier Science L td.