基于改进LightGBM的农机服务备件配置预测方法  

Research on Agricultural Spare Parts Forecasting Based on Improved LightGBM

作　　者：温彦博 王卓[1,2] 白晓平[1,2] Wen Yanbo;Wang Zhuo;Bai Xiaoping(Shenyang Institute of Automation Chinese Academy of Sciences,Shenyang 110000,China;Institutes for Robotics and Intelligent Manufacturing,Chinese Academy of Sciences,Shenyang 110169,China;University of Chinese Academy of Sciences,Beijing 100049,China)

机构地区：[1]中国科学院沈阳自动化研究所,沈阳110000 [2]中国科学院机器人与智能制造创新研究院,沈阳110169 [3]中国科学院大学,北京100049

出　　处：《农机化研究》2024年第4期7-14,共8页Journal of Agricultural Mechanization Research

基　　金：国家重点研发计划项目(2020YFB1709603-1)。

摘　　要：针对农机服务网点中服务备件配置预测不准确导致农机备件资源浪费的问题,根据农机在服务网点的作业情况,提出了一种基于改进LightGBM的农机服务备件配置预测方法。首先,确定了农机作业环境信息、服务点信息以及备件信息三大维度内的多个特征;然后,验证了影响农机服务资源备件量的主要影响因素;接着,基于LightGBM建立了农机服务资源备件预测模型;最后,为了提高模型的精度和速度,通过PSO优化算法对Light-GBM农机服务资源预测模型进行改进,达到了更好的预测结果。实验结果表明:与随机森林、XGBoost等算法相比,LightGBM模型有更好的效果,RMSE值为27.67;通过PSO的超参数调优,LightGBM备件预测的精确性更进一步提高,RMSE值为24.74,能够较为准确地预测农机服务资源在服务网点的备件需求。In view of the current agricultural machinery service network resources distribution and the problem of spare parts waste of resources,this paper proposed a prediction method of spare parts allocation of agricultural machinery service based on improved LightGBM according to the operation of agricultural machinery in the service network.This paper first determines the agricultural machinery operation environment information,service information,and multiple characteristics of spare parts information in three dimensions,then based on PSO-LightGBM agricultural machinery spare parts service resources prediction model is established.To evaluate the effectiveness,we also compared our method with other machine learning methods such as(Logistic Regression,Random Forest,and XGBoost).LightGBM model has a better effect with RMSE value of 27.67.Moreover,The accuracy of LightGBM spare parts prediction is further improved by PSO super parameter tuning,and the RMSE value is 24.74,which can more accurately predict the spare parts demand of agricultural machinery service resources in service outlets.

关 键 词：农机服务 备件预测 LightGBM 机器学习 

分 类 号：S232.8[农业科学—农业机械化工程] TP311.52[农业科学—农业工程]