Özet:

Abstract The problem of text documents classification based on fuzzy probabilistic neural network in real time mode is considered. A different number of classes, which may include such documents, can be allocated in an array of text documents. It is assumed that the data classes can have an n-dimensional space of different shape and mutually overlap. The architecture of the multlayer adaptive fuzzy probabilistic neural network, which allow to solve the problem of classification in sequential mode as new data become available, is.proposed. An algorithm for training the multilayer adaptive fuzzy probabilistic neural network is proposed, and the problem of classification is solved on the basis of the proposed architecture in terms of intersecting classes, which allows to determine the belonging a single instance of a text document to different classes with varying degrees of probability. Classifying neural network architecture characterized by simple numerical implementation and high speed training, and is designed to handle large data sets, characterized by the feature vectors of high dimension. The proposed neural network and its learning method designed to work in conditions of overlapping classes, differing both the form and size. References Specht D. F. Probabilistic neural networks / D. F. Specht // Neural Networks. – 1990. – Vol. 3 (1). – P. 109–118. 2. Бодянский Е. В. Семантическое аннотирование текстовых документов с использованием модифицированной вероятностной нейроной сети / Е. В. Бодянский, О. В. Шубкина // Системные технологии. – Днепропетровск, 2011. – Вып. 4 (75). – С. 48–55. 3. Bodyanskiy Ye. Semantic annotation of text documents using modified probabilistic neural network / Ye. Bodyanskiy, O. Shubkina// Intelligent Data Acquisition and Advanced Computing Systems: Technology and Applications: 6th IEEE Intеrnational Conferences, Prague, 15–17 September 2011: – Prague: Czech Technical University In Prague, 2011. – P. 328–331. 4. Bodyanskiy Ye. Semantic annotation of text documents using evolving neural network based on principle «Neurons at Data Points» / Ye. Bodyanskiy, O. Shubkina // Workshop on Inductive Modelling «IWIM 2011» : 4th Interational Conference, Zhukyn-Kyiv, 4–10 July 2011: Kyiv: IRTC ITS, 2011. – P. 31–37. 5. Bodyanskiy Ye. A learning probabilistic neural network with fuzzy inference / Ye. Bodyanskiy, Ye. Gorshkov, V. Kolodyazhniy, J. Wernstedt // Artificial Neural Nets and Genetic Algorithms «ICANNGA 2003» : 6th International Conference, Roanne, France April 23-25 April 2003 : proceedings. – Wien : Springer-Verlag, 2003. – P. 13–17. 6. Bodyanskiy Ye. Resource-allocating probabilistic neuro-fuzzy network / Ye. Bodyanskiy, Ye. Gorshkov, V. Kolodyazhniy // European Union Society for Fuzzy Logic and Technology «EUSFLAT 2003» : 3rd Internetional Conference, Zittau : proceedings. – Zittau : University of Applied Sciences at Zittau/ Goerlitz, 2003. – P. 392-395. 7. Bodyanskiy Ye. Probabilistic neuro-fuzzy network with nonconventional activation functions / Ye. Bodyanskiy, Ye. Gorshkov, V. Kolodyazhniy, J. Wernstedt // Knowledge-Based Intelligent Information and Engineering Systems : 7th International Conference KES 2003, Oxford, 3–5 September 2003 : proceedings. – Berlin-Heidelberg-New York : Springer, 2003. – P. 973–979. – (Lecture Notes in Computer Science, Vol. 2774). 8. Бодянский Е. В. Классификация текстовых документов с помощью нечеткой вероятностной нейронной сети / Е. В. Бодянский, Н. В. Рябова, О. В. Золотухин // Восточно-европейский журнал передовых технологий – 2011. – № 6/2 (54). – С.16–18. 9. Zahirniak D. R. Pattern recognition using radial basis function network / D. R. Zahirniak, R. Chapman, S. K. Rogers, B. W. Suter, M. Kabriski, V. Pyatti // Aerospace Application of Artificial Intelligence: 6 International Conference, 5–8 June 1990 : proceedings. – Dayton : Ohio, 1990. – P. 249–260.

Anahtar Kelimeler: