Models and Applications of Cooperated Analytical and Heuristical Approaches in a HOSVD-Based Framework

DOI: 10.5176/2251-1679_CGAT13.35

Authors: Andra ́s Ro ̈vid, Pe ́ter Va ́rlaki

Abstract:

The paper introduces various applications of the higher order singular value decomposition (HOSVD) based canonical forms. First of all it is shown, how linear parameter varying (LPV) systems as well as complex logistical processes may be identified by using the above tool.

Further- more its application in the field of image processing especially for image filtering and scaling will be described. Finally the possibilities of its application in the field of neural networks focusing on multilayer perceptrons (MLP) will be introduced.

Keywords: System Identification, Modeling, HOSVD, Ap- proximation