Title:	Computational Performance of the ParametersEstimation in Extreme Seeking EntropyAlgorithm
Authors:	Vrba, Jan Mareš, Jan
Citation:	2020 International Conference on Applied Electronics: Pilsen, 8th – 9h September 2020, Czech Republic.
Issue Date:	2020
Publisher:	Západočeská univerzita v Plzni
Document type:	conferenceObject konferenční příspěvek
URI:	http://hdl.handle.net/11025/39931
ISBN:	978-80-261-0891-7 (Print) 978-80-261-0892-4 (Online)
ISSN:	1803-7232 (Print) 1805-9597 (Online)
Keywords:	zpracování signálu;adaptivní systémy;adaptivní algoritmy;detekce novinky;zobecněná Paretova distribuce
Keywords in different language:	signal processing;adaptive systems;adaptive algorithms;novelty detection;generalized Pareto distribution
Abstract in different language:	This paper is dedicated to the evaluation ofthe computational time performance of the algorithmsthat estimate the parameters of the generalized Paretodistribution, namely Method of Moments, Maximumlikelihood estimator and Quasi-maximum likelihood al-gorithms. The generalized Pareto distribution is utilizedby the Extreme Seeking Entropy algorithm to detectnovelty in data. The algorithm is evaluating the weightincrements of the simple adaptive filter that are obtainedvia incrementally learning algorithm. The computationaltime performance is examined in the experiment withthe detection of step-change parameters of the signalgenerator. Its output contains also additive Gaussiannoise.
Rights:	© Západočeská univerzita v Plzni
Appears in Collections:	Applied Electronics 2020 Applied Electronics 2020

Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.