2D spiral pattern recognition with possibilistic measures

摘要: The main task for a 2D spiral recognition algorithm is to learn discriminate between data distributed on two distinct strands in the x-y plane. This problem of critical importance since it incorporates temporal characteristics often found real-time applications, i.e. coils with time. Previous work this benchmark has witnessed poor results statistical methods such as discriminant analysis and tedious procedures better neural networks. paper describes fuzzy approach which outperforms previous terms rate speed recognition. presents new validation test sets. show that possible solve relatively small amount time (up 100% correct classification set; 77.2% cross-validation using leave-one-out method).