What is the importance of selecting features for non-technical losses identification?

dc.contributorUniversidade Estadual Paulista (UNESP)
dc.creatorRamos, Caio C. O.
dc.creatorPapa, João Paulo
dc.creatorSouza, André N.
dc.creatorChiachia, Giovani
dc.creatorFalcão, Alexandre X.
dc.date2014-05-27T11:25:57Z
dc.date2016-10-25T18:34:16Z
dc.date2014-05-27T11:25:57Z
dc.date2016-10-25T18:34:16Z
dc.date2011-08-02
dc.date.accessioned2017-04-06T01:51:41Z
dc.date.available2017-04-06T01:51:41Z
dc.identifierProceedings - IEEE International Symposium on Circuits and Systems, p. 1045-1048.
dc.identifier0271-4310
dc.identifierhttp://hdl.handle.net/11449/72586
dc.identifierhttp://acervodigital.unesp.br/handle/11449/72586
dc.identifier10.1109/ISCAS.2011.5937748
dc.identifier2-s2.0-79960865826
dc.identifierhttp://dx.doi.org/10.1109/ISCAS.2011.5937748
dc.identifier.urihttp://repositorioslatinoamericanos.uchile.cl/handle/2250/893442
dc.descriptionAlthough non-technical losses automatic identification has been massively studied, the problem of selecting the most representative features in order to boost the identification accuracy has not attracted much attention in this context. In this paper, we focus on this problem applying a novel feature selection algorithm based on Particle Swarm Optimization and Optimum-Path Forest. The results demonstrated that this method can improve the classification accuracy of possible frauds up to 49% in some datasets composed by industrial and commercial profiles. © 2011 IEEE.
dc.languageeng
dc.relationProceedings - IEEE International Symposium on Circuits and Systems
dc.rightsinfo:eu-repo/semantics/closedAccess
dc.subjectAutomatic identification
dc.subjectClassification accuracy
dc.subjectData sets
dc.subjectFeature selection algorithm
dc.subjectIdentification accuracy
dc.subjectNon-technical loss
dc.subjectAutomation
dc.subjectClassification (of information)
dc.subjectParticle swarm optimization (PSO)
dc.subjectFeature extraction
dc.titleWhat is the importance of selecting features for non-technical losses identification?
dc.typeOtro


Este ítem pertenece a la siguiente institución