A Nonparametric Algorithm for Data Preprocessing and Modeling Multidimensional Objects with Delay
Ekaterina Chzhan
Ekaterina Chzhan*, Assistant Professor, School of Space and Information Technology, Siberian Federal University, Krasnoyarsk, Russian Federation.
Manuscript received on November 22, 2020. | Revised Manuscript received on November 25, 2020. | Manuscript published on December 30, 2020. | PP: 22-25 | Volume-10 Issue-2, December 2020. | Retrieval Number: 100.1/ijeat.A19301010120 | DOI: 10.35940/ijeat.A1930.1210220
Open Access | Ethics and Policies | Cite | Mendeley
© The Authors. Blue Eyes Intelligence Engineering and Sciences Publication (BEIESP). This is an open access article under the CC BY-NC-ND license (http://creativecommons.org/licenses/by-nc-nd/4.0/)
Abstract: The paper devotes to modeling tasks of multidimensional inertialess objects with delay (MIOD). The description of identification scheme of MIOD is determined. The proposed identification scheme includes not only blocks of a process and a model but also a data preprocessing block to improve modelling accuracy. A new method of data preprocessing which includes outliers detection and sparsity filling is proposed. It allows generating new training samples based on initial data that is obtained by measurement of input and output variables of the process. A software package is developed to conduct computer experiments. The results of the study show that the proposed algorithms are universal and can be applied to simulate various objects that are described with liner, nonlinear algebraic and nonlinear transcendental mathematical equations. Computational experiments have shown satisfactory accuracy of the algorithms. Proposed algorithms can be used in modeling and control tasks for inertialess objects in various areas of industry such as metallurgy, petrochemicals and etc.
Keywords: Non-parametric algorithm, multidimensional process, data analysis, data preprocessing.
Scope of the Article: Data Analytics