The fourier series model for predicting sapflow density flux based on treetalker monitoring system

The development and application of smart technologies in various fields is increasing every year. Different monitoring systems and sensors generate a large amount of data sets which allows to solve various tasks on data prediction and classification. This paper deals with data sets generated by a new tree monitoring system TreeTalker© which evaluates in particular the sap flow density flux describing water transport in trees. The main task consists in prediction of the values of this characteristic which reflects the tree life state based only on observable air temperature during the predictable time interval and subsequent classification of trees according to some prespecified classes. The Fourier series based model is used to fit the data sets with periodic patterns. The multivariate regression model defines the functional dependencies between sap flow density and temperature time series. The paper shows that Fourier coefficients can be successfully used as elements of the feature vectors required to solve different classification problems. Artificial multilayer neural networks are used as classifiers. The quality of the developed model for prediction and classification is verified by numerous numerical examples. © Springer Nature Switzerland AG 2020.

Authors
Efrosinin D. 1, 2 , Kochetkova I. 2, 3 , Stepanova N.4 , Yarovslavtsev A. 2, 5 , Samouylov K. 2, 3 , Valentini R. 2, 6
Language
English
Pages
198-209
Status
Published
Volume
12526 LNCS
Year
2020
Organizations
  • 1 Johannes Kepler University Linz, Altenbergerstrasse 69, Linz, 4040, Austria
  • 2 Peoples’ Friendship University of Russia (RUDN University), 6 Miklukho-Maklaya St, Moscow, 117198, Russian Federation
  • 3 Federal Research Center “Computer Science and Control” of the Russian Academy of Sciences, 44-2 Vavilova St, Moscow, 119333, Russian Federation
  • 4 V.A. Trapeznikov Institute of Control Sciences of RAS, Profsoyuznaya St., 65, Moscow, 117997, Russian Federation
  • 5 LAMP, Russian Timiryazev State Agrarian University, 49 Timiryazevskaya st, Moscow, 127550, Russian Federation
  • 6 Tuscia University, Via S.M. in Gradi n.4, Viterbo, 01100, Italy
Keywords
Fourier series; Multivariate linear regression; Neural network; Time series prediction and classification; Tree monitoring
Date of creation
20.04.2021
Date of change
20.04.2021
Short link
https://repository.rudn.ru/en/records/article/record/71799/
Share

Other records

Paramonov A., Tonkikh E., Koucheryavy A., Tatarnikova T.M.
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 12525 LNCS. 2020. P. 307-316
Melnikov S.Y., Samouylov K.E.
Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics). Vol. 12526 LNCS. 2020. P. 259-269