基于改进Filter和随机森林的应急物资供应商评价方法研究
Research on Emergency Material Supplier Evaluation Method Based on Improved Filter and Random Forest
刘志勇 1耿琳1
作者信息
- 1. 河北工业职业技术大学,河北石家庄 050091
- 折叠
摘要
针对应急物资供应商评价时因主观因素影响较大、存在冗余指标等情况,导致供应商评价分类效果不佳的问题,提出基于改进Filter和随机森林的方法改善应急物资供应商评价分类预测效果的思路.利用Filter中拉普拉斯分值算法的高效计算能力进行初步特征优化,并在此过程中引入互信息法消除冗余特征,构建随机森林分类器,以期在消除主观因素影响的同时,降低冗余指标造成的不利后果.实验数据显示,该方法可以明显提升应急物资供应商评价时的分类预测效果,为应急供应链管理的快速构建和高效运作提供可靠依据.
Abstract
In response to the problem of poor classification performance in emergency material supplier evaluation due to subjective factors and redundant indicators,a method based on improved Filter and random forest is proposed to improve the classification and prediction performance of emergency material supplier evaluation.Utilizing the efficient computing power of the Laplace score algorithm in Filter for preliminary feature optimization,and introducing mutual information method to eliminate redundant features in this process,a random forest classifier is constructed to reduce the adverse effects caused by redundant indicators while eliminating subjective factors.Experimental data shows that this method can significantly improve the classification and prediction performance of emergency material supplier evaluation,providing a reliable basis for the rapid construction and efficient operation of emergency supply chain management.
关键词
改进Filter/互信息法/随机森林/应急物资供应商/评价Key words
improve Filter/mutual information method/random forest/emergency material supplier/evaluation引用本文复制引用
出版年
2024
NETL
NSTL
万方数据