Ramos, Caio C. O.Papa, João Paulo [UNESP]Souza, André N. [UNESP]Chiachia, GiovaniFalcão, Alexandre X.2014-05-272014-05-272011-08-02Proceedings - IEEE International Symposium on Circuits and Systems, p. 1045-1048.0271-4310http://hdl.handle.net/11449/72586Although 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.1045-1048engAutomatic identificationClassification accuracyData setsFeature selection algorithmIdentification accuracyNon-technical lossAutomationClassification (of information)Particle swarm optimization (PSO)Feature extractionTrabalho apresentado em evento10.1109/ISCAS.2011.5937748Acesso aberto2-s2.0-799608658268212775960494686