基于组合分类策略的个人信用风险评估研究

随着个人贷款的需求日渐旺盛,评估贷款人的信用好坏对金融机构规避风险尤为重要。文章提出了基于组合分类策略的个人信用风险评估模型,该模型引入了决策分值的概念,选取K最近邻(K-nearest neighbor,KNN)、随机森林(random forest,RF)、决策树(decision tree,DT)和支持向量机(support vector machine SVM)等常用的信用评估分类算法作为基分类器,分别从稳定性和准确性2个维度计算每个基分类器的决策分值,并根据组合后的决策分值评估贷款人信用好坏,最后在3个真实数据集上验证了该模型的有效性。

OPTIMAL HIERARCHY STRUCTURES FOR MULTI-ATTRIBUTE-CRITERIA DECISIONS

A problem of a hierarchy structure optimization is considered.Hierarchical structures arewidely used in the Analytic Hierarchy Process,conjoint analysis,and various other methods of multiplecriteria decision making.The problem consists in finding a structure that needs a minimum number ofpair comparisons for a given total number of the alternatives.For an optimal hierarchy,the minimumefforts are needed for eliciting data and synthesizing the local preferences across the hierarchy to getthe global priorities or utilities.Special estimation techniques are developed and numerical simulationsperformed.Analytical and numerical results suggest optimal ways of priority evaluations for practicalmanagerial decisions in a complex environment.