Decomposition and the principle of interaction prediction in hierarchical structure of learning algorithm of ANN

• Autorzy: Płaczek S.
• Konferencja: Computer Applications in Electrical Engineering (20-21.04.2015 ; Poznań, Polska)
• Języki publikacji: EN

Abstrakty

EN

For the most popular ANN structure with one hidden layer, decomposition is done into two sub-networks. These sub-networks form the first level of the hierarchical structure. On the second level, the coordinator is working with its own target function. In the hierarchical systems theory three coordination strategies are defined. For the ANN learning algorithm the most appropriate is the coordination by the principle of interaction prediction. Implementing an off-line algorithm in all sub-networks makes the process of weight coefficient modification more stable. In the article, the quality and quantity characteristics of a coordination algorithm and the result of the learning algorithm for all sub-networks are shown. Consequently, the primary ANN achieves the global minimum during the learning process.

Słowa kluczowe

EN

artificial neural network; hierarchy; decomposition; coordination; coordination principle

• Wydawca: Wydawnictwo Politechniki Poznańskiej
• Czasopismo: Poznan University of Technology Academic Journals. Electrical Engineering
• Rocznik: 2015
• Tom: No. 84
• Strony: 113--120
• Opis fizyczny: Bibliogr. 6 poz., rys.

Twórcy

autor

Płaczek S.

• Akademia Finansów i Biznesu Vistula

Bibliografia

• [1] M.D. Mesarocic, D. Macko and Y. Takahara, Theory of hierarchical multilevel systems, Academic Press, New York and London 1970.
• [2] Ch.M. Bishop, Pattern Recognition and Machine Learning, Springer Science + Business Media, LLC 2006.
• [3] W. Findeisen, J. Szymanowski, A. Wierzbicki, Teoria i metody obliczeniowe optymalizacji, Państwowe Wydawnictwo Naukowe, Warsaw 1977.
• [4] Zeng-Guang Hou, Madan M. Gupta, Peter N. Nikiforuk, Min Tan and Long Cheng, "A Recurrent Neural Network for Hierarchical Control of Interconnected Dynamic Systems", IEEE Transactions on Neural Networks, Vol. 18, No. 2, March 2007.
• [5] S. Placzek, “A two-level on-line learning algorithm of Artificial Neural Network with forward connections”. Proc. Science and Information Conference 2014, London, ISBN: 978-0-9893193-1-7.
• [6] S. Placzek, B. Adhikari, "Analysis of Multilayer Neural Network with Direct Connection Cross-forward Connection”, CS&P Conference 2013, The University of Warsaw, Warsaw 2013.