Novas Estratégias para Treinamento de Least Squares Support Vector Machines

Bernardo Penna Resende de CarvalhoAntônio P. Braga

We present in this work two new training strategies for the LS-SVMs, in order to eliminate their greatest drawback when comparing to SVMs, the inexistence of the support vectors’ automatic detection. Least Squares Support Vector Machines (LS-SVMs) were created in 1999, corresponding to a modified version of Support Vector Machines (SVMs), developed in 1992. The main characteristic of the LS-SVMs is the low computational complexity comparing to the SVMs, without quality loss in the solution, because the principles that both have been based are the same. In this work, we considered the strategies Ada

Caso o link acima esteja inválido, faça uma busca pelo texto completo na Web: Buscar na Web

Biblioteca Digital Brasileira de Computação - Contato:
     Mantida por: