讲座时间：2022年4月26日（周二）14:00-15:30

腾讯会议号：365 648 395

主讲人：寇纲

讲座题目：Profit-and risk-driven credit scoringunder parameter uncertainty: a multi-objective approach

讲座语言：中文

讲座摘要：

Profit-drivenmachine learning models and profit-based performance measures have been widelyused in credit scoring. When assessing the performance of a machine learningmodel for credit scoring, previous research typically assumes that the cost andbenefit parameters, and their distributional information are available.However, in reality, these parameters and their distributions are often notexactly known. This paper considers the parameter uncertainty in thedevelopment of credit scoring models, and the estimation of profits and risksgenerated by employing those models. We propose a novel profit-based metric-theworst-case expected minimum cost (WEMC)-to estimate the profit of creditscoring models with uncertain parameters. Furthermore, we introduce theworst-case conditional value-at-risk measure (WCVaR) to measure the loss incurredfrom employing a classification model in credit scoring during thedeterioration of cost parameters. A multi-objective feature-selection frameworkgrounded on WEMC and WCVaR is then presented for model development. We employtwelve credit scoring datasets from multiple countries to compare the proposedmethods, with feature selection methods that use metrics including the areaunder the receiver operating characteristic curve, the minimum cost, and theexpected minimum cost as selection criteria. The results suggest that theproposed methods outperform other feature-selection methods in terms of costand risk performance metrics.

主讲人简介：

寇纲，教授，博士生导师。现任西南财经大学大数据研究院院长、工商管理学院执行院长、长江学者特聘教授、国家杰出青年科学基金获得者、全国MBA教育指导委员会委员、国务院享受政府特殊津贴专家。