基于PCA-BP神经网络的羊肉价格预测研究

Prediction of mutton price based on PCA-BP neural network

价格波动受到多种因素的影响,准确的价格预测对市场稳定具有重要的作用。为了更好地服务于市场经济,笔者选取苏尼特羊为研究对象,提出了基于PCA和BP神经网络结合的价格预测方法,首先对影响羊肉价格的指标数据进行PCA降维,选取主成分特征最大的因子为BP神经网络的输入层,羊肉价格作为输出层建立预测模型,将预测结果与已有的苏尼特羊肉价格进行比对验证。结果表明:与标准的BP神经网络相比,结合PCA降维的BP模型具有更小的拟合误差和更高的精度。

Price fluctuation is affected by many factors,and accurate price forecast plays an important role in market stability.This author selected Sunite sheep as the research object and proposed a price prediction method based on the combination of PCA and BP neural network.Firstly,the PCA dimensionality reduction was applied to the index data affecting the price of mutton.The factor with the largest principal component,and the mutton price were selected as the input layer and the output layer of BP neural network respectively to establish the prediction model.The prediction result was compared with the current price of Sunite mutton.The results indicated that the BP model combined with PCA dimensionality reduction has smaller fitting error and higher precision compared with the standard BP neural network.