This paper aims to presenting a new algorithm to classify symbolic data. The input data for the learning step is a set of symbolic objects described by symbolic interval (or set-valued) variables. At the end of the learning step, each group is represented by a (modal) symbolic object which is described by symbolic histogram (or bar-diagram) variables. The assignment of a new observation to a group is based on a dissimilarity function which measures the difference in content and in position between them. The difference in position is measured by a context free component whereas the difference in content is measured by a context dependent component. To show the usefulness of this modal symbolic pattern classifier, a particular kind of simulated images is classified according to this approach.
I documenti in IRIS sono protetti da copyright e tutti i diritti sono riservati, salvo diversa indicazione.