Application Research of Graph Neural Networks in the Financial Risk Control

Zhongbao Yu; Jiaqi Zhang; Xin Qi; Chao Chen

doi:doi:10.11648/j.mcs.20220706.14

| Peer-Reviewed

Application Research of Graph Neural Networks in the Financial Risk Control

Zhongbao Yu, Jiaqi Zhang, Xin Qi, Chao Chen

Published in Mathematics and Computer Science (Volume 7, Issue 6)

Received: 8 April 2022 Accepted: 26 December 2022 Published: 28 December 2022

Views: Downloads:

Download PDF

Share This Article

Twitter
Linked In
Facebook

Abstract

Combining deep learning with graph data, the method applied to learning tasks on association relationships is collectively referred to as Graph neural network (GNN). This paper mainly studies the application of GNN in the financial risk control. With the enterprise customer network graph, this paper designs a credit rating model based on GNN, an implicit relationship recognition model, and a fusion model of the two. To reduce duplication and improve model performance, graph pruning method is introduced in the data preprocessing stage, such as entity fusion, relationship normalization, etc. According to the prediction results, the heterogeneous graph credit rating model is better than the homogeneous one. Moreover, the suspicious relations detected by the implicit relation recognition model can be complementary to the heterogeneous graph credit rating model, which will improve the model performance. The model of this paper can be applied not only in the financial risk control, but also can provide a reference for other fields. In response to external public opinion information, the credit rating model label is effectively supplemented, and the heterogeneous graph credit rating model is used to learn related topology information, redefine the credit rating of related enterprises, leading to discover related high-risk enterprises, and achieve the purpose of risk control. This is an advantage that traditional machine learning methods do not have.

Published in	Mathematics and Computer Science (Volume 7, Issue 6)
DOI	10.11648/j.mcs.20220706.14
Page(s)	124-129
Creative Commons	This is an Open Access article, distributed under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/), which permits unrestricted use, distribution and reproduction in any medium or format, provided the original work is properly cited.
Copyright	Copyright © The Author(s), 2022. Published by Science Publishing Group

Keywords

Graph Neural Networks, Heterogeneous Graph, Graph Pruning, Risk Control

References

[1]	Kaoudi Z, Lorenzo A, Markl V. Towards Loosely-Coupling Knowledge Graph Embeddings and Ontology-based Reasoning [J]. 2022.
[2]	Gong F, Wang M, Wang H, et al. SMR: Medical Knowledge Graph Embedding for Safe Medicine Recommendation [J]. Big Data Research, 2021, 23: 100174.
[3]	Patel A, Pai S S, Rajamohan H R, et al. Finding Novel Links in COVID-19 Knowledge Graph Using Graph Embedding Techniques [C]// Smoky Mountains Computational Sciences and Engineering Conference. Springer, Cham, 2022.
[4]	Bulla M, Hillebrand L, Lübbering, Max, et al. Knowledge Graph Based Question Answering System for Financial Securities [J]. 2021.
[5]	Liu Y, Zhang M, Ma C, et al. Graph neural network [J]. 2020.
[6]	Chen Z, Ma T, Y Wang. When Does A Spectral Graph Neural Network Fail in Node Classification? [J]. 2022.
[7]	Hanik M, Demirta M A, Gharsallaoui M A, et al. Predicting Cognitive Scores With Graph Neural Networks Through Sample Selection Learning [J]. 2021.
[8]	Kurshan E, Shen H. Graph Computing for Financial Crime and Fraud Detection: Trends, Challenges and Outlook [J]. International Journal of Semantic Computing, 2020, 14 (04): 565-589.
[9]	Zhang L. Knowledge graph theory and structural parsing [J]. university of twente, 2002.
[10]	Gainullina A N, Shalyto A A, Sergushichev A A. Method of the Joint Clustering in Network and Correlation Spaces [J]. Modeling and Analysis of Information Systems, 2020, 27 (2): 180-193.
[11]	Gainullina A N, Shalyto A A, Sergushichev A A. Method for Joint Clustering in Graph and Correlation Spaces [J]. Automatic Control and Computer Sciences, 2022, 55 (7): 647-657.
[12]	Zhongbao Y, Fangming S, Zuyuan Z. Researches for more reliable arrangement graphs in multiprocessor computer system [J]. Applied Mathematics and Computation, 2019, 363: 124611.
[13]	Zhongbao Y, Fangming S. Approximation Algorithm of Arrangement Graph Reliability in Parallel System [J]. Journal of East China University of Science and Technology, 2020, 46 (6): 838-842.
[14]	Ye W, Huang Z, Hong Y, et al. Graph Neural Diffusion Networks for Semi-supervised Learning [J]. 2022.
[15]	Hamilton W L, Ying R, Leskovec J. Inductive Representation Learning on Large Graphs [J]. 2017.
[16]	Chang L, Branco P. Graph-based Solutions with Residuals for Intrusion Detection: the Modified E-GraphSAGE and E-ResGAT Algorithms [J]. 2021.
[17]	Chen Z, Deng Q, Zhao Z, et al. Energy consumption prediction of cold source system based on GraphSAGE. 2021.
[18]	Hajibabaee P, Malekzadeh M, Heidari M, et al. An Empirical Study of the GraphSAGE and Word2vec Algorithms for Graph Multiclass Classification [C]// 2021 IEEE 12th Annual Information Technology, Electronics and Mobile Communication Conference (IEMCON). IEEE, 2021.
[19]	Schlichtkrull M, Kipf T N, Bloem P, et al. Modeling Relational Data with Graph Convolutional Networks [C]. European Semantic Web Conference. Springer, Cham, 2018.
[20]	Sun D, Ma L, Ding Z, et al. An Attention-Driven Multi-label Image Classification with Semantic Embedding and Graph Convolutional Networks [J]. 2022.
[21]	Y Yao, Joe-Wong C. FedGCN: Convergence and Communication Tradeoffs in Federated Training of Graph Convolutional Networks [J]. 2022.
[22]	Niu H, Haitao H E, Feng J, et al. Knowledge Graph Completion Based on GCN of Multi-Information Fusion and High-Dimensional Structure Analysis Weight [J]. Chinese Journal of Electronics, 2022.
[23]	J Gao, X Liu, Chen Y, et al. MHGCN: Multiview Highway Graph Convolutional Network for Cross-Lingual Entity Alignment [J]. Tsinghua Science and Technology, 2022, 27 (4): 719-728.

Cite This Article

Plain Text BibTeX RIS

APA Style

Zhongbao Yu, Jiaqi Zhang, Xin Qi, Chao Chen. (2022). Application Research of Graph Neural Networks in the Financial Risk Control. Mathematics and Computer Science, 7(6), 124-129. https://doi.org/10.11648/j.mcs.20220706.14

Copy | Download

ACS Style

Zhongbao Yu; Jiaqi Zhang; Xin Qi; Chao Chen. Application Research of Graph Neural Networks in the Financial Risk Control. Math. Comput. Sci. 2022, 7(6), 124-129. doi: 10.11648/j.mcs.20220706.14

Copy | Download

AMA Style

Zhongbao Yu, Jiaqi Zhang, Xin Qi, Chao Chen. Application Research of Graph Neural Networks in the Financial Risk Control. Math Comput Sci. 2022;7(6):124-129. doi: 10.11648/j.mcs.20220706.14

Copy | Download

@article{10.11648/j.mcs.20220706.14,
  author = {Zhongbao Yu and Jiaqi Zhang and Xin Qi and Chao Chen},
  title = {Application Research of Graph Neural Networks in the Financial Risk Control},
  journal = {Mathematics and Computer Science},
  volume = {7},
  number = {6},
  pages = {124-129},
  doi = {10.11648/j.mcs.20220706.14},
  url = {https://doi.org/10.11648/j.mcs.20220706.14},
  eprint = {https://article.sciencepublishinggroup.com/pdf/10.11648.j.mcs.20220706.14},
  abstract = {Combining deep learning with graph data, the method applied to learning tasks on association relationships is collectively referred to as Graph neural network (GNN). This paper mainly studies the application of GNN in the financial risk control. With the enterprise customer network graph, this paper designs a credit rating model based on GNN, an implicit relationship recognition model, and a fusion model of the two. To reduce duplication and improve model performance, graph pruning method is introduced in the data preprocessing stage, such as entity fusion, relationship normalization, etc. According to the prediction results, the heterogeneous graph credit rating model is better than the homogeneous one. Moreover, the suspicious relations detected by the implicit relation recognition model can be complementary to the heterogeneous graph credit rating model, which will improve the model performance. The model of this paper can be applied not only in the financial risk control, but also can provide a reference for other fields. In response to external public opinion information, the credit rating model label is effectively supplemented, and the heterogeneous graph credit rating model is used to learn related topology information, redefine the credit rating of related enterprises, leading to discover related high-risk enterprises, and achieve the purpose of risk control. This is an advantage that traditional machine learning methods do not have.},
 year = {2022}
}

Copy | Download

TY  - JOUR
T1  - Application Research of Graph Neural Networks in the Financial Risk Control
AU  - Zhongbao Yu
AU  - Jiaqi Zhang
AU  - Xin Qi
AU  - Chao Chen
Y1  - 2022/12/28
PY  - 2022
N1  - https://doi.org/10.11648/j.mcs.20220706.14
DO  - 10.11648/j.mcs.20220706.14
T2  - Mathematics and Computer Science
JF  - Mathematics and Computer Science
JO  - Mathematics and Computer Science
SP  - 124
EP  - 129
PB  - Science Publishing Group
SN  - 2575-6028
UR  - https://doi.org/10.11648/j.mcs.20220706.14
AB  - Combining deep learning with graph data, the method applied to learning tasks on association relationships is collectively referred to as Graph neural network (GNN). This paper mainly studies the application of GNN in the financial risk control. With the enterprise customer network graph, this paper designs a credit rating model based on GNN, an implicit relationship recognition model, and a fusion model of the two. To reduce duplication and improve model performance, graph pruning method is introduced in the data preprocessing stage, such as entity fusion, relationship normalization, etc. According to the prediction results, the heterogeneous graph credit rating model is better than the homogeneous one. Moreover, the suspicious relations detected by the implicit relation recognition model can be complementary to the heterogeneous graph credit rating model, which will improve the model performance. The model of this paper can be applied not only in the financial risk control, but also can provide a reference for other fields. In response to external public opinion information, the credit rating model label is effectively supplemented, and the heterogeneous graph credit rating model is used to learn related topology information, redefine the credit rating of related enterprises, leading to discover related high-risk enterprises, and achieve the purpose of risk control. This is an advantage that traditional machine learning methods do not have.
VL  - 7
IS  - 6
ER  -

Copy | Download

Author Information

Zhongbao Yu

Data management and Application Department, Bank of Shanghai, Shanghai, China
Jiaqi Zhang

Data management and Application Department, Bank of Shanghai, Shanghai, China
Xin Qi

Data management and Application Department, Bank of Shanghai, Shanghai, China
Chao Chen

Data management and Application Department, Bank of Shanghai, Shanghai, China

Download PDF

Sections

Plain Text BibTeX RIS

APA Style

Zhongbao Yu, Jiaqi Zhang, Xin Qi, Chao Chen. (2022). Application Research of Graph Neural Networks in the Financial Risk Control. Mathematics and Computer Science, 7(6), 124-129. https://doi.org/10.11648/j.mcs.20220706.14

Copy | Download

ACS Style

Zhongbao Yu; Jiaqi Zhang; Xin Qi; Chao Chen. Application Research of Graph Neural Networks in the Financial Risk Control. Math. Comput. Sci. 2022, 7(6), 124-129. doi: 10.11648/j.mcs.20220706.14

Copy | Download

AMA Style

Zhongbao Yu, Jiaqi Zhang, Xin Qi, Chao Chen. Application Research of Graph Neural Networks in the Financial Risk Control. Math Comput Sci. 2022;7(6):124-129. doi: 10.11648/j.mcs.20220706.14

Copy | Download

@article{10.11648/j.mcs.20220706.14,
  author = {Zhongbao Yu and Jiaqi Zhang and Xin Qi and Chao Chen},
  title = {Application Research of Graph Neural Networks in the Financial Risk Control},
  journal = {Mathematics and Computer Science},
  volume = {7},
  number = {6},
  pages = {124-129},
  doi = {10.11648/j.mcs.20220706.14},
  url = {https://doi.org/10.11648/j.mcs.20220706.14},
  eprint = {https://article.sciencepublishinggroup.com/pdf/10.11648.j.mcs.20220706.14},
  abstract = {Combining deep learning with graph data, the method applied to learning tasks on association relationships is collectively referred to as Graph neural network (GNN). This paper mainly studies the application of GNN in the financial risk control. With the enterprise customer network graph, this paper designs a credit rating model based on GNN, an implicit relationship recognition model, and a fusion model of the two. To reduce duplication and improve model performance, graph pruning method is introduced in the data preprocessing stage, such as entity fusion, relationship normalization, etc. According to the prediction results, the heterogeneous graph credit rating model is better than the homogeneous one. Moreover, the suspicious relations detected by the implicit relation recognition model can be complementary to the heterogeneous graph credit rating model, which will improve the model performance. The model of this paper can be applied not only in the financial risk control, but also can provide a reference for other fields. In response to external public opinion information, the credit rating model label is effectively supplemented, and the heterogeneous graph credit rating model is used to learn related topology information, redefine the credit rating of related enterprises, leading to discover related high-risk enterprises, and achieve the purpose of risk control. This is an advantage that traditional machine learning methods do not have.},
 year = {2022}
}

Copy | Download

TY  - JOUR
T1  - Application Research of Graph Neural Networks in the Financial Risk Control
AU  - Zhongbao Yu
AU  - Jiaqi Zhang
AU  - Xin Qi
AU  - Chao Chen
Y1  - 2022/12/28
PY  - 2022
N1  - https://doi.org/10.11648/j.mcs.20220706.14
DO  - 10.11648/j.mcs.20220706.14
T2  - Mathematics and Computer Science
JF  - Mathematics and Computer Science
JO  - Mathematics and Computer Science
SP  - 124
EP  - 129
PB  - Science Publishing Group
SN  - 2575-6028
UR  - https://doi.org/10.11648/j.mcs.20220706.14
AB  - Combining deep learning with graph data, the method applied to learning tasks on association relationships is collectively referred to as Graph neural network (GNN). This paper mainly studies the application of GNN in the financial risk control. With the enterprise customer network graph, this paper designs a credit rating model based on GNN, an implicit relationship recognition model, and a fusion model of the two. To reduce duplication and improve model performance, graph pruning method is introduced in the data preprocessing stage, such as entity fusion, relationship normalization, etc. According to the prediction results, the heterogeneous graph credit rating model is better than the homogeneous one. Moreover, the suspicious relations detected by the implicit relation recognition model can be complementary to the heterogeneous graph credit rating model, which will improve the model performance. The model of this paper can be applied not only in the financial risk control, but also can provide a reference for other fields. In response to external public opinion information, the credit rating model label is effectively supplemented, and the heterogeneous graph credit rating model is used to learn related topology information, redefine the credit rating of related enterprises, leading to discover related high-risk enterprises, and achieve the purpose of risk control. This is an advantage that traditional machine learning methods do not have.
VL  - 7
IS  - 6
ER  -

Copy | Download