Classification of time realizations using machine learning recognition of recurrence plots

Кіріченко, Л. О.; Радівілова, Т. А.; Зінченко, П. П.

Публікація:
Classification of time realizations using machine learning recognition of recurrence plots

dc.contributor.author	Кіріченко, Л. О.
dc.contributor.author	Радівілова, Т. А.
dc.contributor.author	Зінченко, П. П.
dc.date.accessioned	2021-06-10T20:17:13Z
dc.date.available	2021-06-10T20:17:13Z
dc.date.issued	2021
dc.description.abstract	In the article,the machine learning classification of time realizations using the recurrence plot visualization is considered. Every time realization is converted intoa matrix of recurrence states and it is presented as a black-and-white image. The resulting images of realizations are classified using deep neural networks. A deep residual neural network is used as an image classifier. The binary classifi-cation of EEG realizationsis carried out. The result of the binary classification is the detection of an epileptic seizure. The data for the experiment are records of brain activity containing 178 values. The results of the studyshow that the consi-dered method has a high classification accuracy. The proposed classification ap-proach can be readily used in practice.	uk_UA
dc.identifier.citation	Kirichenko L. Classification of time realizations using machine learning recognition of recurrence plots / L. Kirichenko, P. Zinchenko, T. Radivilova // Advances in Intelligent Systems and Computing. – 2021. – P. 687–696.	uk_UA
dc.identifier.uri	https://openarchive.nure.ua/handle/document/16446
dc.language.iso	en	uk_UA
dc.publisher	AISC	uk_UA
dc.subject	machine learning classification	uk_UA
dc.subject	time series classification	uk_UA
dc.subject	recurrence plot	uk_UA
dc.subject	EEG realizations	uk_UA
dc.subject	deep residual networks	uk_UA
dc.title	Classification of time realizations using machine learning recognition of recurrence plots	uk_UA
dc.type	Article	uk_UA
dspace.entity.type	Publication