基于PSO-BP与RBF神经网络的蔬菜价格组合预测

Vegetables Price Combination Forecasting Based on PSO-BP and RBF Neural Network

为准确预测蔬菜市场价格走势,现选取海南省儋州市2012—2015年117组青椒旬零售价格及相关因素的旬价格为样本数据,其中100组作为训练数据,17组数据作为测试数据,分别建立基于粒子群算法优化BP神经网络的蔬菜价格预测模型和基于RBF神经网络的蔬菜价格预测模型,并在这2种模型的基础上建立蔬菜价格的线性组合预测模型。结果表明:2种单项预测模型在蔬菜价格预测上的应用效果都较好,且在不同评价指标上各有优势。将这2种模型的预测结果进行线性组合,可以使各单项模型优势互补,拟合效果明显优于各单项预测模型。

In order to predict vegetables price accurately,117 groups 2012—2015green pepper and related factors price in danzhou city were selected as the sample data,of which 100 groups were training data and 17 groups were test data,the PSO-BP forecasting model and the RBF network forecasting model concerning vegetables retail price were set up separately,and then the linear combination forecasting model was set up on the basis of these two models.The results indicated that these two single forecasting models＇ effect was well,and two models had their own advantages in different evaluation index.The linear combination of the prediction results of these two models could make single model＇s advantage complementary,whose fitting effect was better than the single forecasting model.