%0 Conference Proceedings
%4 sid.inpe.br/sibgrapi/2016/09.01.13.41
%2 sid.inpe.br/sibgrapi/2016/09.01.13.41.44
%T Aprendizado incremental e classe-incremental por meio da atualização de árvores geradoras em florestas de caminhos ótimos
%D 2016
%A Riva, Mateus,
%A Campos, Teófilo de,
%A Ponti, Moacir,
%@affiliation ICMC/Universidade de São Paulo
%@affiliation CVSSP/University of Surrey, FGA/Universidade de Brasília
%@affiliation ICMC/Universidade de São Paulo
%E Aliaga, Daniel G.,
%E Davis, Larry S.,
%E Farias, Ricardo C.,
%E Fernandes, Leandro A. F.,
%E Gibson, Stuart J.,
%E Giraldi, Gilson A.,
%E Gois, João Paulo,
%E Maciel, Anderson,
%E Menotti, David,
%E Miranda, Paulo A. V.,
%E Musse, Soraia,
%E Namikawa, Laercio,
%E Pamplona, Mauricio,
%E Papa, João Paulo,
%E Santos, Jefersson dos,
%E Schwartz, William Robson,
%E Thomaz, Carlos E.,
%B Conference on Graphics, Patterns and Images, 29 (SIBGRAPI)
%C São José dos Campos, SP, Brazil
%8 4-7 Oct. 2016
%I Sociedade Brasileira de Computação
%J Porto Alegre
%S Proceedings
%K aprendizado incremental, minimum spanning trees, optimum-path forests.
%X Class-incremental algorithms, where there is the need to update classification models with data that emerges over time, are important in many applications. We report an algorithm for updating optimum-path forests capable of including a new instance, maintaining the properties of the optimum-path trees. In addition to a proof demonstrating that the algorithm maintains the structure of the trees in linear time, the experimental evidence shows the applicability of the method, which starting from a limited model, and after the inclusion of multiple instances, is able to achieve the same accuracy of a classifier trained with the full training set.
%@language pt
%3 OPF_CI___SIBGRAPI___WUW__Portugues_(1).pdf