摘要
钢板表面颜色是判定钢板表面耐蚀性能的重要指标,提前预测钢板表面颜色可以为控制钢板表面耐蚀性能提供指导。针对单一模型预测精度较低的情况,提出一种基于Stacking的组合模型。该模型采用两层模式,第一层使用支持向量机、随机森林、GBDT等七个个体学习器作为初级学习器,第二层使用XGBoost作为次级学习器。使用该方法对钢板表面颜色进行预测,结果表明,基于Stacking的组合模型与单一模型相比,在多个性能指标上取得了明显的提升。
The surface color of steel plate is an important pointer for determining the corrosion resistance of steel plate surface.Predicting the surface color of steel plate in advance can provide guidance for controlling the corrosion resistance of steel plate surface.In view of the low prediction accuracy of single model,this paper proposes a combined model based on Stacking.This method uses a two-layer model,the first layer integrated seven individual learners such as SVM,Random Forest,GBDT as primary learners,and the second layer used XGBoost as a secondary learner.Using this method to predict the surface color of steel plate,compared with the single model,the results show that the combined model based on Stacking has achieved significant improvement in multiple performance indicators.
作者
刘媛媛
赵希庆
Liu Yuanyuan;Zhao Xiqing(Department of Mathematics and Information Technology,Yuncheng University,Yuncheng,Shanxi 044000,China;Department of Mechanical and Electrical Engineering,Yuncheng University)
出处
《计算机时代》
2020年第8期65-68,共4页
Computer Era
基金
运城学院博士科研启动项目(YQ-2019003)。