Formation classification models of undetermined data by means of procedures reduction and kappa statistic

Authors

  • Skalozub Vladyslav
  • Horiachkin Vadim
  • Terlitskyi Ihor
  • Dudnyk Ilya

DOI:

https://doi.org/10.34185/1562-9945-5-148-2023-13

Keywords:

classification, reliable models, dimensionality of space, fuzzy values, CF(A) certainty factors, modified Hamming network, reduction procedure, Cohen's kappa statistic, Ukrainian-language texts, author's definition, computer simulation.

Abstract

The article is devoted to the development of mathematical models for the classi-fication of uncertain data represented by fuzzy values and certainty factors CF(A). Diagnostic pattern formation procedures use modified Hamming networks (MHN), as well as reduction methods and Cohen's kappa statistics. At the same time, the limit-ing dimensions and composition of the parameters of the classification model are de-termined, which ensure the established probabilistic requirements for the reliability of the calculation results. The model space reduction procedure and the structure of the software complex for diagnosing uncertain data are presented. An example of a clas-sification model based on fuzzy data is the task of identifying the authors of Ukrain-ian-language texts. The classification task for data in CF(A) format corresponds to candidate selection. The results of the numerical modeling made it possible to estab-lish the effectiveness, reliability and efficiency of the proposed procedures for the formation of reliable classification models with uncertain data.

References

Skalozub V.V., Goryachkin V.M., Klymenko I.V., Terletsky I.A., Terlenko A.P. Dos-lidzhennia protsedur merezhi khemminha dlia upravlinnia servisnymy systema-my pry netochno vyznachenykh i pryrodomovnykh danykh. 2022. No. 3-4 (99-100). P. 33–47. DOI: https://doi.org/10.15802/stp2022/276411

Velykoivanenko, H.I. (2018). Otsiniuvannia rivnia ekonomichnoi bezpeky na pidgrunti vidstani Khemminha. Retrieved from https://core.ac.uk/download/pdf/197269753.pdf (in Ukrainian)

Vasilev, V.I. (1998). Induktsiya i reduktsiya v problemakh ekstrapolyatsii. Cybernetics and Computer Engineering, 116, 65-81. (in Russian).

Kolesnyk A. S., Khairova N. F. . Obgruntuvannia vykorystannia statystyky kappa Koena v eksperymentalnykh doslidzhenniakh NLP Text Mining Cybernetics and sys-tem analysis. Vol. 58, No. 2. 2022. P. 143–153.

Li Min Fu, Shortliffe E. H. The application of certainty factors to neural computing for rule discovery. IEEE Transactions on Neural Networks. 2000. Vol. 11. Iss. 3. P. 647–657. DOI: https://doi.org/10.1109/72.846736

Skalozub V.V., Goryachkin V.M., Terletskyi I.A. Bahatoparametrychni intelektualni protsedury diahnostuvannia za nepovnymy i zburenymy danymy// Logistics and transport safety: Problems and prospects of development in the context of analysis of modern challenges and threats: materials of reports II International scientific and practical conference. — Dnipro: Serednyak T.K., 2023.

P. 42-47.

Shinkarenko V. I., Demidovych I. M. Determination of signs of authorship of natu-ral language texts. Artificial Intelligence. 2018. No. 3. P. 27–35.

Richard A. Brualdi. Combinatorial matrix classes. — Cambridge: Cambridge Uni-versity Press, 2006. — (Encyclopedia of Mathematics and Its Applications). — ISBN 0-521-86565-4

Leszek Rutkowski Metody i techniki sztucznej inteligentcji. Naukove PWN, Warsaw, 2005. – 520 p.

Haykin S. Neural networks: A Comprehensive Foundation. Prentice hall: New Jer-sey, 1999.1103 p.

Downloads

Published

2023-12-19