METHOD AND SYSTEM FOR EVALUATING THE CLASS OF TEST DATA IN A LARGE-DIME SION DATA SPACE

摘要

The present invention relates to a method and a system for evaluating the class of test data in an original metric system, each data belonging to at least one class in which a plurality of data is grouped. The method comprises a step of graphically representing the space organization of a learning data set of the original space in a metric representation space, a joined membership level making it possible to know if any two data of the learning set belong to a same class. The method also comprises a step of relating the test data to projections of data of the learning set, the most likely class of the test data being the class of the projections of the data of the learning set in relation to the test data. The invention can be used for assisting in the discrimination and recognition shapes.