Possibilistic Kohonen maps
Résumé
A Possibilistic Kohonen network based essentially on two fuzzy measures (the possibility and the necessity degrees) is proposed in this paper. This type of network can take into account the imperfection of the information elements in the objects of the training set (the imprecision, the uncertainty, and the missing values) when processing the signals. A concrete numeric example is given to clarify and to easily explain the proposed algorithm.