基于数学建模方法的矿石加工质量预测研究  

Research on Quality Prediction of Ore Processing Based on Mathematical Modeling Methods

作　　者：周湘辉 许凯 ZHOU Xiang-hui;XU Kai(School of Mathematics and Statistics,Anhui Normal University,Wuhu 241002,China)

机构地区：[1]安徽师范大学数学与统计学院,安徽芜湖241002

出　　处：《安徽师范大学学报（自然科学版）》2024年第5期401-406,共6页Journal of Anhui Normal University(Natural Science)

基　　金：国家自然科学基金研究项目(12271005);安徽省自然科学基金研究项目(2308085Y06);安徽省质量工程基金研究项目(2020jyxm0649).

摘　　要：本文基于数学建模方法对矿石加工质量指标进行了预测研究。应用两个Excel附件数据表,将数据分为三种情形,运用五类模型(线性回归、随机森林、决策树、XGBoost、GBDT模型)对相关数据进行了不同维度的数据预测.为了得到模型的拟合精度,应用了决策系数法对五类模型的预测效果进行了定量分析,最终选择决策树与GBDT模型分别对矿石加工质量指标与矿石加工合格率进行了预测研究.本研究通过实验验证了所提出模型的有效性和可行性,对于提升工业生产效率和产品质量具有一定的实际应用价值。A prediction problem on the quality indicators of ore processing is studied by mathematical modeling methods.Two data tables of Excel attachment are used to divide the data into three types.Five models(linear re-gression,random forest,decision tree,XGBoost,GBDT model)are used to predict the relevant data from dif-ferent dimensions.In order to obtain the fitting accuracy of the model,the method of decision coefficient is ap-plied to quantitatively analyze the prediction effect of the five models.Naturally,the decision tree and GBDT models are selected to predict the quality indicators of ore processing and the qualification rate of ore processing,respectively.This study verifies the effectiveness and feasibility of the proposed models via experiments,which has certain practical application value for improving industrial production efficiency and product quality.

关 键 词：矿石加工 质量与合格率 决策树 GBDT算法 

分 类 号：O29[理学—应用数学]