Classification flow system errors with a hybrid modification of neural network Wang-Mendel
Автор: Soldatova Olga Petrovna, Pudikova Ekaterina Mikhailovna
Журнал: Известия Самарского научного центра Российской академии наук @izvestiya-ssc
Рубрика: Интеллектуальные информационные системы
Статья в выпуске: 4-2 т.16, 2014 года.
Бесплатный доступ
The article explores the application of the proposed authors hybrid modification of neural fuzzy multiple outputs Wang-Mendel network to solve a problem of error messages flow classification in the software system. Describes the properties of the message flow represented in natural language, methodology of training set formation, which based on real error messages stream. Describes the network learning algorithms and the efficiency of hybrid network for solving the problem.
Message flows in natural language training set, fuzzy neural production network, wang-mendel, self-organizing layer, learning algorithm classification
Короткий адрес: https://sciup.org/148203209
IDR: 148203209
