Performance evaluation for engineering project management of particle swarm optimization based on least squares support vector machines

Dong Qiao-Ting, Geng Li-Yan an

Abstract

As for the limitation of using cross validation method to choose the parameters of least squares support vector machines (LSSVM), this paper proposes a new classified model which combines adaptive particle swarm optimization (APSO) algorithm with LSSVM. The new model uses APSO algorithm to select optimal parameters for LSSVM. According to the analysis of the management performance evaluation for engineering project, we conclude that LSSVM-APSO has better evaluation performance than LSSVM which bases on cross validation method. On searching for the optimal parameters of LSSVM, APSO algorithm is obviously faster than that by cross validation method.