融合气象因子的苏尼特羊肉价格预测方法研究  

作　　者：喻夏 周李涌[1] 

机构地区：[1]内蒙古科技大学,内蒙古包头014010

出　　处：《智慧农业导刊》2022年第13期37-40,共4页JOURNAL OF SMART AGRICULTURE

基　　金：内蒙古自治区研究生教育教学改革研究与实践项目(YJG20191012710);内蒙古自治区科技重大专项(2019ZD025);内蒙古自然科学基金(2019MS06021)。

摘　　要：畜牧业的牲畜价格会由多种因素影响,准确预测牲畜的价格,能更好地指导农户生产、销售,实现利益的最大化。文章以苏尼特羊并结合气象因子为研究,提出由核PCA及动态贝叶斯网络(DBN)相结合的方法预测价格。该模型首先采用核PCA降维处理数据,选择相关性最大的主要因素作为动态贝叶斯网络的输入参数,以输出参数为苏尼特羊肉价格搭建预测模型。然后将所得结果与苏尼特羊肉的实际价格进行比较。最终比照显示,与标准动态贝叶斯网络模型对照,动态贝叶斯网络模型结合核PCA具有更好的拟合性和更准确的精度。Livestock prices in animal husbandry are affected by a variety of factors. Accurate prediction of livestock prices can better guide farmers’ production and sales and maximize benefits. In this paper, based on Sunit sheep and meteorological factors, a method combining kernel PCA and Dynamic Bayesian Network is proposed to predict prices. The model first uses kernel PCA to reduce the dimensionality of the data, selects the main factor with the greatest correlation as the input parameter of the Dynamic Bayesian Network, and uses the output parameter to build a prediction model for Sunit mutton price. The results obtained are then compared with the actual price of Sunit sheep. The final comparison shows that,compared with the standard Dynamic Bayesian Network model, the Dynamic Bayesian Network model combined with kernel PCA has better fitting and more accurate accuracy.

关 键 词：核PCA 价格预测 苏尼特羊 动态贝叶斯网络 

分 类 号：S126[农业科学—农业基础科学]