基于GA-XGBoost算法的河南省粮食产量预测研究

Grain yield prediction based on GA-XGBoost algorithm in Henan Province

粮食问题关乎国家命运,是国民经济发展基础中的基础。粮食产量的变化直接关系到我国的粮食安全和农业结构的优化调整。为提高河南省粮食产量预测的精度和效率,对河南省粮食产量等相关数据进行归纳分析,利用皮尔逊(Pearson)相关性影响分析确定主要影响河南省粮食产量的因素。针对XGBoost模型容易过拟合、预测不精准的问题,引入遗传算法(GA)对其学习率、树的深度等进行优化,以更准确地预测河南省粮食产量。仿真结果表明:相比于传统的XGBoost模型,GA-XGBoost模型具有更高的预测精度,RMSE仅为0.034。因此,GA-XGBoost预测模型可以对粮食产量进行更为准确的预测。

The food issue is related to the fate of the country,and the food issue is the foundation of the national economic de-velopment foundation.The change of grain output is directly related to the food security and the optimization and adjustment of agri-cultural structure in our country.In order to improve the accuracy and efficiency of grain production forecasting in Henan Province,relevant data such as grain production in Henan Province were summarized and analyzed,and the main factors affecting grain pro-duction in Henan Province were determined by Pearson correlation impact analysis.Aiming at the problem that XGBoost model is easy to overfit and inaccurate in prediction,genetic algorithm(GA)was introduced to optimize its learning rate and tree depth,so as to predict grain yield in Henan Province more accurately.The simulation results show that compared with the traditional XGBoost model,GA-XGBoost model has higher prediction accuracy,RMSE is only 0.034.Therefore,GA-XGBoost forecasting model can make more accurate prediction of grain yield.