Radiotekhnika
Publishing house Radiotekhnika

"Publishing house Radiotekhnika":
scientific and technical literature.
Books and journals of publishing houses: IPRZHR, RS-PRESS, SCIENCE-PRESS


Тел.: +7 (495) 625-9241

 

Adaptive forecasting model of time series

Keywords:

S.A. Tonoyan - Ph.D. (Eng.), Associate Professor, Department «Information processing and control systems»,  Bauman Moscow State Technical University
E-mail: tonoyansl@mail.ru
M.V. Chernenkiy - Associate Professor, Department «Information processing and control systems», Bauman Moscow State Technical University
E-mail: chernen @ bmstu.ru
A.S. Tonoyan - Assistent, Department «Information processing and control systems», Bauman Moscow State Technical University
E-mail: tonoyanas@mail.ru


Prediction of defects and timely assessment of technical condition beyond allows you to improve the readiness and effec-tiveness of the functioning of the technical system as a whole. The process of early detection of defects and evaluation of technical conditions of complex systems in the current time and extrapolating underlines the relevance of the methods build prediction models.
The success of the task determines the adequacy of the forecast model. The choice of method and the build forecasting model is largely overstated from the problem domain and source data. One of the areas of forecasting, this time series, the scope of which is quite broad.
A comparative analysis of existing methods of forecasting time series classification simplifies the selection of an ade-quate method, and its adaptation. Information model-based forecast is more accurate, because it takes into account the data properties. Therefore, can adequately describe the dynamics of change processes in the form of time series and play with patterns in the data.
Considering the controlled parameters, characterizing the state of the system as a function of time, you can solve the problem of predicting changes its status through extrapolation.
Adaptive prediction model belongs to a class of information models, it takes into account the trend and significance of time series data using weights that are adapted in the following steps.
On the basis of developed and implemented an algorithm in the form of a software module conducted an experiment. The data obtained indicate that the accuracy of the adaptive forecasting model on average 20% higher than considered.

References:
  1. Brillindzher D.R. Vremennы́e ryady'. Obrabotka danny'x i teoriya. M.: Mir. 1980. 532 s.
  2. Bendat Dzh., Pirsol A. Prikladnoj analiz danny'x. M.: Mir. 1989. 540 s.
  3. Paklin N.B., Oreshkov V.I. Biznes analitika: ot danny'x k znaniyam. SPb.: Piter. 2013. 704 s.
  4. Barsegyan A.A. Analiz danny'x i proczessov: Ucheb. posobie. SPb.: BXV-Peterburg. 2009. 512 s.
  5. Tonoyan S.A. Razrabotka struktury' avtomatizirovannoj sistemy' upravleniya texnicheskim sostoyaniem ob''ektov aviaczionnoj texniki i prognoziruyushhego kontrolya: Avtoref. ... kand. texn. nauk. M. 1985. 16 s.
  6. Chety'rkin E.M. Staticheskie metody' prognozirovaniya. M. 1975. 291 s.
  7. Tonoyan S.A., Saraev D.V. Temporal'ny'e modeli bazy' danny'x i ix svojstva // Inzhenerny'j zhurnal: nauka i innovaczii. E'lektronnoe nauchno-texnicheskoe izdanie. 2014. Vy'p. № 12 (36).
  8. Vlasova E.A. Ryady'. M.: MGTU im. N.E'. Baumana. 2006. 616 s.
  9. Dejper N., Smit G. Prikladnoj regressionny'j analiz. M.: Finansy' i Statistika. 1987. 343 s.
  10. Mejndoland Dzh. Vy'chislitel'ny'e algoritmy' v prikladnoj statistike. M.: Finansy' i Statistika. 1988. 349 s.
  11. Baldin A.V., Tonoyan S.A., Eliseev D.V. Analiz izby'tochnosti xraneniya temporal'ny'x danny'x sredstvami relyaczionny'x SUBD // Inzhenerny'j zhurnal: nauka i innovaczii. 2014. Vy'p. № 4(24). http://engjournal.ru/articles/1273/1273.pdf
  12. Kende'l M. Vremennы́e ryady'. M.: Finansy' i statistika. 1981. 329 s.
June 24, 2020
May 29, 2020

© Издательство «РАДИОТЕХНИКА», 2004-2017            Тел.: (495) 625-9241                   Designed by [SWAP]Studio