一种基于粒子群优化算法的组合预测模型

A Combination Prediction Model Based on Particle Swarm Optimization

本文首先分析了若干传统的预测方法,提出了一种组合预测模型,在该模型中利用加权系数对各种预测方法进行组合,集成不同来源的预测结果,从不同的侧面反映整个预测过程,力图使预测结果更加精确。在各种预测方法加权系数的确定上,利用PSO快速全局优化的特点,可以减少试算的盲目性,提高模型预测的准确性。

Several traditional prediction methods are analyzed and a combination prediction model is proposed in this paper. In the proposed model,weight-coefficients are used to combine various prediction methods,and integrate the prediction results with different sources, so as to reflect the whole prediction process from different aspects and to make the prediction results more accurate. PSO, which has the characteristics of fast global optimization, is used to determine the weight-coefficients for various prediction methods. This approach can reduce the blindness of search and increase the prediction precision of the model.