An improved neural approach of Sammon projection algorithm

According to the pattern recognition theory, a very important stage into a classification chain is represented by the feature selection. Although in literature a lot of feature selection techniques are indicated, one of the most important methods as application area is focused on Sammon projection algorithm use. Consequently, in this paper an improved neural approach of Sammon mapping is described. In addition, using a real HRR classification task, a comparison as performance level of the proposed method with other well-known neural versions of Sammon projection technique is also included.