Artificial intelligence

Abstract: The actual problem of automation of construction and analysis of recognition and diagnostic models based on neural networks was solved. The set of criteria for comparing of neural network models and synthesis methods was proposed. It contains in addition to the traditionally used time and accuracy criteria the criteria characterizing the logical transparency. Computational experiments on the proposed criteria for comparison of neural networks were conducted by solving practical problems.

Keywords: neural network pattern recognition, diagnosis, model analysis, comparison criterion.

References:

1. Kruglov V. V. Iskusstvennye nejronnye seti: Teorija i praktika / V. V. Kruglov, V. V. Borisov. – M.:Gorjachaja linija-Telekom, 2001. – 382 p.
2. Rambidi N. G. Biomolekuljarnye nejrosetevye ustrojstva / N. G. Rambidi. – M.: IPRZhR, 2002.– 224 c.
3. Intellektual'nye sredstva diagnostiki i prognozirovanija nadezhnosti aviadvigatelej: mono-grafija / [V. I. Dubrovin,S. A. Subbotin, A. V. Boguslaev, V. K. Jacenko]. – Zaporozh'e: OAO "Motor-Sich", 2003. – 279 p.
4. Intellektual'nye informacionnye tehnologii proektirovanija avtomatizirovannyh sistem diagnostirovanija iraspoznavanija obrazov : monografija / [S. A. Subbotin, An. A. Olejnik, E. A. Gofman, S. A. Zajcev, Al.A. Olejnik] ; pod red. S. A. Subbotina. – Har'kov : OOO «Kom-panija Smit», 2012. – 317 p.
5. Progressivnye tehnologii modelirovanija, optimizacii i intellektual'noj avtomatizacii jeta-pov zhiznennogocikla aviacionnyh dvigatelej : monografija / [A. V. Boguslaev, Al. A. Olejnik, An. A. Olejnik, D. V.Pavlenko, S. A. Subbotin] ; pod red. D. V. Pavlenko, S. A. Subbotina. – Zaporozh'e: OAO "Motor Sich",2009. – 468 p.