Normalization Issues in Granular Evidence-Based Adaptive Network ANGIE

International Conference on Computational Intelligence for Modelling, Control and Automation (CIMCA 2003) 2003
Aleksandrs Vališevskis, Arkādijs Borisovs

Granular evidence-based adaptive network ANGIE (adaptive network for granular information processing) is considered in this paper. The network described can be used in fuzzy-evidence-based decision support systems. The architecture and learning algorithm underlying ANGIE is presented. The proposed learning procedure helps in solving a task that can be related to sensitivity analysis in decision aid models. However, normalization procedure should be applied to network parameters to avoid inconsistency. This paper considers different approaches towards normalization. The effectiveness and accuracy of different approaches are compared.

Keywords
adaptive network, information granularity, fuzzy evidence, decision support systems, sensitivity analysis
Hyperlink
http://alephfiles.rtu.lv/TUA01/000022342_e.pdf

Vališevskis, A., Borisovs, A. Normalization Issues in Granular Evidence-Based Adaptive Network ANGIE. In: International Conference on Computational Intelligence for Modelling, Control and Automation (CIMCA 2003), Austria, Vienna, 12-14 February, 2003. Canberra: University of Canberra, 2003, pp.117-124.

Publication language
English (en)