基于蜻蜓算法优化随机森林回归的粮食产量预测研究  

Study on grain yield prediction using random forest regression based on dragonfly algorithm

作　　者：仝晓春[1] 周玲[1] Tong Xiaochun;Zhou Ling(Xi'an Siyuan University,Shaanxi Xi’an 710038,China)

机构地区：[1]西安思源学院,陕西西安710038

出　　处：《现代科学仪器》2023年第1期186-192,共7页Modern Scientific Instruments

基　　金：陕西省高校工程研究中心研究项目(编号:20200318)。

摘　　要：为有效提升粮食产量预报的精确性,利用RFR算法进行粮食生产预测的问题,建立了基于DA-RFR技术的粮食生产预测模型、叶节点的样本数以及候选特征的子集等关键参数影响较大的问题。粮食产量作为DA-RFR的输出,建立DA-RFR粮食预测模型。与GA-RFR、PSO-RFR和RFR相比,DA-RFR能够有效提高粮食产量预测的精度,其训练集和预测集的预测精度分别高达2.14%和2.03%,为保障国家粮食安全、进一步提高规划的科学性,为行政机关进行政策制定工作提供有力支持。In order to effectively improve the accuracy of grain yield forecast,RFR algorithm was used to forecast grain production.The prediction model of grain production based on DA-RFR technology,sample number of leaf nodes,subset of candidate characteristics and other key parameters were established.As the output of DA-RFR,the DA-RFR grain forecasting model was established.Compared with GA-RFR,PSO-RFR and RFR,DA-RFR can effectively improve the accuracy of grain yield prediction.The accuracy of the training set and the prediction set are as high as 2.14%and 2.03%,respectively.It provides strong support for ensuring national food security,further improving the scientific planning,and providing strong support for the administrative organs to make policy.

关 键 词：粮食产量 遗传算法 随机森林 蜻蜓算法 粒子群算法 

分 类 号：F32[经济管理—产业经济]