TBU Publications
Repository of TBU Publications

Lenses classification by means of pseudo neural networks - Two approaches

DSpace Repository

Show simple item record


dc.title Lenses classification by means of pseudo neural networks - Two approaches en
dc.contributor.author Komínková Oplatková, Zuzana
dc.contributor.author Šenkeřík, Roman
dc.relation.ispartof MENDEL 2014
dc.identifier.issn 1803-3814 OCLC, Ulrich, Sherpa/RoMEO, JCR
dc.identifier.isbn 9788021449848
dc.date.issued 2014
utb.relation.volume 2014-January
utb.relation.issue January
dc.citation.spage 397
dc.citation.epage 402
dc.event.title 20th International Conference on Soft Computing: Evolutionary Computation, Genetic Programming, Swarm Intelligence, Fuzzy Logic, Neural Networks, Fractals, Bayesian Methods, MENDEL 2014
dc.event.location Brno
utb.event.state-en Czech Republic
utb.event.state-cs Česká republika
dc.event.sdate 2014-06-25
dc.event.edate 2014-06-27
dc.type conferenceObject
dc.language.iso en
dc.publisher Brno University of Technology
dc.subject Analytic programming en
dc.subject Artificial neural networks en
dc.subject Classifiers en
dc.subject Evolutionary computation en
dc.subject Optimization en
dc.subject Pseudo neural networks en
dc.subject Symbolic regression en
dc.description.abstract This research deals with a novel approach to classification. This paper deals with a synthesis of a complex structure, which serves as a classifier. This structure is similar to classical artificial neural net therefore the name pseudo neural network is used. The proposed method for classifier structure synthesis utilizes Analytic Programming (AP) as the tool of the evolutionary symbolic regression. AP synthesizes a whole structure of the relation between inputs and output. Classical artificial neural networks, where a relation between inputs and outputs is based on the mathematical transfer functions and optimized numerical weights, were an inspiration for this work. The paper shows two approaches - continues classification with one output node and classical approach with binary classification and more output nodes. Lenses data (one of benchmarks for classifiers) was used for testing of the proposed method. For experimentation, Differential Evolution for the main procedure and also for meta-evolution version of analytic programming was used. en
utb.faculty Faculty of Applied Informatics
dc.identifier.uri http://hdl.handle.net/10563/1005268
utb.identifier.obdid 43871863
utb.identifier.scopus 2-s2.0-84938078069
utb.source d-scopus
dc.date.accessioned 2015-08-28T12:04:58Z
dc.date.available 2015-08-28T12:04:58Z
utb.contributor.internalauthor Komínková Oplatková, Zuzana
utb.contributor.internalauthor Šenkeřík, Roman
Find Full text

Files in this item

Show simple item record