数据挖掘评估中融合多目标决策算法应用研究  

Application Research of Multi-objective Decision Algorithm in Evaluation of Data Mining

作　　者：陈辉[1] CHEN Hui(Huainan Vocational and Technical College,Huainan 232001,Anhui,China)

机构地区：[1]淮南职业技术学院,安徽淮南232001

出　　处：《贵阳学院学报（自然科学版）》2020年第3期5-9,共5页Journal of Guiyang University：Natural Sciences

基　　金：安徽省教育厅质量工程项目“淮南职业技术学院安徽华联商厦有限责任公司电子商务创业实践教育基地”(项目编号:2018sjjd108)。

摘　　要：机器学习中评估模型的性能是当前较为热门的方向,尤其是在数据挖掘过程中如何评估、高效地获取最优决策算法,以此来达到最佳的决策分类效率。文章提出了融合目标值和专家权重的多目标决策算法,专家权重采用定量化方式,且在判断群决策矩阵中使用最小二乘法来优化AFP的权重计算过程。以某银行信用卡信息来验证文章算法,对经典的分类算法进行多目标决策,并应用评估模型来获取得到最佳的分类算法。The performance of the evaluation model in machine learning is currently the hottest direction,especially in the data mining process,how to evaluate the efficient acquisition to achieve the optimal decision algorithm,in order to obtain the best decision classification efficiency.In this paper,a multi-objective decision-making algorithm combining target value and expert weight is proposed.The expert weight is quantified,and the least squares method is used in the judgment group decision matrix to optimize the weight calculation process of AFP.The bank's credit card information is used to verify the algorithm,the multi-objective decision algorithm is applied to the classical classification algorithm,and the evaluation model is applied to obtain the best classification algorithm.

关 键 词：评估 数据挖掘 最小二乘法 多目标决策 

分 类 号：TP309[自动化与计算机技术—计算机系统结构] O225[自动化与计算机技术—计算机科学与技术]