Deep neural networks with L1 and L2 regularization for high dimensional corporate credit risk prediction

Yang, Mei; Lim, Ming K.; Qu, Yingchi; Li, Xingzhi; Ni, Du

doi:10.1016/j.eswa.2022.118873

Cited by 35 publications

(14 citation statements)

References 86 publications

Supporting

Mentioning

Contrasting

Unclassified

Order By: Relevance

“…In addition, the ANN architecture has been expanded to a deep neural network with five hidden layers and hundreds of hidden neurons for each layer. The L1 and L2 regularisation are included to prevent overtraining by adding the penalty to the loss function (Yang et al, 2023). As illustrated in Figure 2B, the predicted capacitance profile (red line) has a similar trend to the experimental capacitance (blue line).…”

Section: Prediction Ability Of Machine Learningmentioning

confidence: 76%

Transition of electrochemical measurement to machine learning in the perspective of two-dimensional materials

2022

View full text Add to dashboard Cite

Two-dimensional materials (e.g. graphene, and transition metal dichalcogenides) have become ubiquitous in electrochemical contexts including energy storage, electrocatalyst, and ion-selective membranes. This is due to its superior electrochemical properties, specifically “capacitance”, which can be referred to the storage ions at the electrolyte/materials interfaces. Experimental work and computational chemistry were carried out in the past decade for solving and improving the understanding of two-dimensional materials; however, these techniques are relatively expensive, complex, and time-consuming. Therefore, we accentuate the future trend of two-dimensional material study with machine learning as the modest alternative. In this perspective, the intrinsic capacitance properties of the two dimension materials were described from an atomic level, explaining the heteroatom doping to a nanoscopic level, showing (basal vs edge capacitance). The studies also extended to the macroscopic level i.e., the flake size of the two-dimensional materials. We then shed more light on the applicability of machine learning coupled with the “fundamental measurement” for solving electrochemistry of two-dimensional materials. The shallow artificial neural network was demonstrated for the prediction of CV curves using the data from size-dependent graphene. In addition, the application of deep neural networks with complicated architecture has also been explored through the prediction of capacitance for heteroatom-doped graphene. This perspective provides a clear background and creates the connection between fundamental measurement and machine learning for understanding the capacitance properties of two-dimensional materials.

show abstract

Section: Prediction Ability Of Machine Learningmentioning

confidence: 76%

Transition of electrochemical measurement to machine learning in the perspective of two-dimensional materials

2022

View full text Add to dashboard Cite

show abstract

“…Although the factors influencing enterprise credit are constantly being explored, most previous studies focused on the optimization and improvement of evaluation methods and paid less attention to empirical analysis (Djeundje et al. , 2021; Kriebel and Stitz, 2022; Yang et al. , 2023).…”

Section: Discussion Implications and Limitationsmentioning

confidence: 99%

“…Furthermore, little research attention has been paid to the empirical analysis of the influence mechanism of soft information on enterprise credit. Most previous studies focused on the optimization and improvement of evaluation methods and paid less attention to empirical analysis (Djeundje et al, 2021;Kriebel and Stitz, 2022;Yang et al, 2023). Wang et al (2022b) explored the role of social and psychologically related-soft information in credit analyses.…”

Section: Literature Review 21 Corporate Credit Behaviormentioning

confidence: 99%

“…Third, few studies have proposed a suitable theoretical framework to understand corporate credit behavior. Yang et al. (2023) proposed an improved deep neural network (DNN) algorithm to evaluate corporate credit risk in a high-dimensional dataset.…”

Section: Introductionmentioning

confidence: 99%

“…Third, few studies have proposed a suitable theoretical framework to understand corporate credit behavior. Yang et al (2023) proposed an improved deep neural network (DNN) algorithm to evaluate corporate credit risk in a high-dimensional dataset. Chen and Chen (2022) developed an approach to evaluate corporate credit ratings by analyzing public opinion of corporations on social media to assist financial institutions in effectively evaluating and controlling corporate risk.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

How social activities affect corporate credit behavior? – the mediating role of network centrality

Chen

2023

IMDS

View full text Add to dashboard Cite

PurposeSocial information is crucial to credit ratings and can improve the accuracy of the traditional credit assessment model. Drawing on the resource-based view (RBV) and social capital theory (SCT), this research explores the relationships between corporate social activities, network centrality and corporate credit behavior.Design/methodology/approachThe authors used social network analysis (SNA) and regression analysis to analyze the data collected from 14,544 enterprises on the Alibaba platform.FindingsThe results indicate that among the four types of social activities, the number of corporate questions and posts shows a positive relationship with credit behavior; while the number of corporate comments has negative relationship with credit behavior. Further, degree and betweenness centralities mediate the relationship between the number of corporate questions, posts and comments with credit behavior.Originality/valueThis study contributes to the literature on non-financial factors (soft information) by exploring the social behavioral factors related to corporate credit. In addition, this study offers a new theoretical lens and reasonable explanations for investigating the relationship between corporate social activities, network centrality and credit behavior from the perspective of the resource-based view, while most studies are predictive and methodological. Moreover, this study provides new insights for platforms to evaluate enterprise credit and for managers to improve credit behavior.

show abstract

Two‐stage credit risk prediction framework based on three‐way decisions with automatic threshold learning

Li,

Sha

2024

Journal of Forecasting

View full text Add to dashboard Cite

Credit risk prediction is a binary classification problem. Using two‐way decisions to classify defaulters may lead to decision errors due to insufficient information. To solve this issue, in addition to identifying borrowers as defaulters and nondefaulters, this paper introduced the delay‐decision mechanism in three‐way decisions, so that records acquiring more information do not make decisions immediately. A two‐stage credit risk prediction framework based on three‐way decisions was proposed to reduce decision risk. In this framework, the decision cost values of three‐way decisions were simplified by analyzing the credit risk prediction, and the expression of threshold calculation was also modified. An optimization objective was built according to the trade‐off between information gain and decision cost, and the particle swarm optimization algorithm was applied to learn the decision thresholds. After adding more supplementary information, the samples in the delayed‐decision region were made further decisions. A dataset from a commercial bank in China was employed to conduct experiments, and the results demonstrated that our proposed method outperformed various base classifiers.

show abstract

Deep neural networks with L1 and L2 regularization for high dimensional corporate credit risk prediction

Cited by 35 publications

References 86 publications

Transition of electrochemical measurement to machine learning in the perspective of two-dimensional materials

Transition of electrochemical measurement to machine learning in the perspective of two-dimensional materials

How social activities affect corporate credit behavior? – the mediating role of network centrality

Two‐stage credit risk prediction framework based on three‐way decisions with automatic threshold learning

Contact Info

Product

Resources

About