15th International Conference on Pattern Recognition (ICPR'00) - Volume 2
Generalized Non-Reducible Descriptors
Barcelona, Spain
September 03-September 08
ISBN: 0-7695-0750-6
This paper provides a generalization of non-reducible descriptors. Non-reducible descriptors are used in supervised pattern recognition problems when the pattern descriptions consist of Boolean variables. This generalization extends the concept of distance between patterns of different classes. A mathematical model to construct generalized non-reducible descriptors, a computational procedure, and numerical examples are discussed.
Index Terms:
Supervised Pattern Recognition, Machine Learning, Descriptors, Non-Reducible Descriptors, Generalized Non-Reducible Descriptors
Citation:
Ventzeslav Valev, Bülent Sankur, Petia Radeva, "Generalized Non-Reducible Descriptors," icpr, vol. 2, pp.2394, 15th International Conference on Pattern Recognition (ICPR'00) - Volume 2, 2000