Monotonic Neural Additive Models: Pursuing Regulated Machine Learning Models for Credit Scoring

被引:4
|
作者
Chen, Dangxing [1 ]
Ye, Weicheng [1 ]
机构
[1] Duke Kunshan Unnivers, Zu Chongzhi Ctr Math & Computat Sci, Kunshan, Jiangsu, Peoples R China
关键词
neural networks; model explainability; fairness; ART CLASSIFICATION ALGORITHMS;
D O I
10.1145/3533271.3561691
中图分类号
F8 [财政、金融];
学科分类号
0202 ;
摘要
The forecasting of credit default risk has been an active research field for several decades. Historically, logistic regression has been used as a major tool due to its compliance with regulatory requirements: transparency, explainability, and fairness. In recent years, researchers have increasingly used complex and advanced machine learning methods to improve prediction accuracy. Even though a machine learning method could potentially improve the model accuracy, it complicates simple logistic regression, deteriorates explainability, and often violates fairness. In the absence of compliance with regulatory requirements, even highly accurate machine learning methods are unlikely to be accepted by companies for credit scoring. In this paper, we introduce a novel class of monotonic neural additive models, which meet regulatory requirements by simplifying neural network architecture and enforcing monotonicity. By utilizing the special architectural features of the neural additive model, the monotonic neural additive model penalizes monotonicity violations effectively. Consequently, the computational cost of training a monotonic neural additive model is similar to that of training a neural additive model, as a free lunch. We demonstrate through empirical results that our new model is as accurate as black-box fully-connected neural networks, providing a highly accurate and regulated machine learning method.
引用
收藏
页码:70 / 78
页数:9
相关论文
共 50 条
  • [21] Credit scoring, augmentation and lean models
    Banasik, J
    Crook, J
    JOURNAL OF THE OPERATIONAL RESEARCH SOCIETY, 2005, 56 (09) : 1072 - 1081
  • [22] An application of hybrid models in credit scoring
    Bonilla, M
    Olmeda, I
    Puertas, R
    FINANCIAL MODELLING, 2000, : 69 - 78
  • [23] Comparison of neuropsychological models of learning with machine learning models in the context of neural networks
    Perus, Mitja
    Elektrotehniski Vestnik/Electrotechnical Review, 1997, 64 (2-3): : 136 - 141
  • [24] JointLIME: An interpretation method for machine learning survival models with endogenous time-varying covariates in credit scoring
    Chen, Yujia
    Calabrese, Raffaella
    Martin-Barragan, Belen
    RISK ANALYSIS, 2024,
  • [25] FROM CREDIT SCORING TO REGULATORY SCORING: COMPARING CREDIT SCORING MODELS FROM A REGULATORY PERSPECTIVE
    Xia, Yufei
    Liao, Zijun
    Xu, Jun
    LI, Yinguo
    TECHNOLOGICAL AND ECONOMIC DEVELOPMENT OF ECONOMY, 2022, 28 (06) : 1954 - 1990
  • [26] On Propagated Scoring for Semisupervised Additive Models
    Culp, Mark
    JOURNAL OF THE AMERICAN STATISTICAL ASSOCIATION, 2011, 106 (493) : 248 - 259
  • [27] Credit scoring using ensemble machine learning
    Yao, Ping
    HIS 2009: 2009 NINTH INTERNATIONAL CONFERENCE ON HYBRID INTELLIGENT SYSTEMS, VOL 3, PROCEEDINGS, 2009, : 244 - 246
  • [28] Credit Risk Analysis Using Machine and Deep Learning Models
    Addo, Peter Martey
    Guegan, Dominique
    Hassani, Bertrand
    RISKS, 2018, 6 (02):
  • [29] Time to Assess Bias in Machine Learning Models for Credit Decisions
    Brotcke, Liming
    JOURNAL OF RISK AND FINANCIAL MANAGEMENT, 2022, 15 (04)
  • [30] Visual analytics for monitoring credit scoring models
    Baldo, Daiane Rodrigues
    Regio, Murilo Santos
    Manssour, Isabel Harb
    INFORMATION VISUALIZATION, 2023, 22 (04) : 340 - 357