果酒品质评价的自适应量子粒子群LS-SVM模型

针对BP神经网络和遗传算法对果酒品质预测存在速度慢和精度低的缺点,建立了一种基于量子行为粒子群算法(QPSO)的最小二乘支持向量机(LS-SVM)的果酒品质预测模型。模型通过引入粒子的进化度和聚合度,动态调整收缩扩张因子,从而实现了算法的动态自适应性。仿真结果表明:基于自适应量子粒子群的LS-SVM果酒品质评价预测模型优于所比较的BP神经网络和最小二乘支持向量机两种模型,具有较好的泛化性能和预测精度。

The Model of Wine Quality Evaluation Based on Adaptive QPSO-LSSVM

Due to the disadvantages of slow speed and low precision of the wine quality evaluation by using BP neural network and Genetic Algorithm, a prediction model of Least Squares Support Vector Machines (LS-SVM) based on quantum behaved particle swarm optimization (QPSO) has been established. By adopting the particle's evolution degree and polymerization degree, contraction expansion factor keeps changing as the particle's evolution degree factor and polymerization degree factor vary in order to realize the dynamic self-adaptive algorithm. The simulations reveal that the AQPSO-LSSVM model of wine quality evaluation is superior to the other two models, BP neural network and Genetic Algorithm. The AQPSO-LSSVM model has better generalization performance and prediction accuracy.