logo

SCIENTIA SINICA Informationis, Volume 51 , Issue 7 : 1068(2021) https://doi.org/10.1360/SSI-2020-0028

Risk assessment for contagion path in complex loan network

More info
  • ReceivedFeb 16, 2020
  • AcceptedJun 15, 2020
  • PublishedJun 23, 2021

Abstract


Funded by

国家重点研发计划(2018AAA0100704)

中国博士后科学基金(2019M651499)


References

[1] Jian M, Xu M. Determinants of the guarantee circles: The case of Chinese listed firms. Pac-Basin Finance J, 2012, 20: 78-100 CrossRef Google Scholar

[2] Meng X F, Tong Y H, Liu X H, et al. Netrating: credit risk evaluation for loan guarantee chain in china. In: Proceedings of Pacific-Asia Workshop on Intelligence and Security Informatics, Korea, 2017. 99--108. Google Scholar

[3] Wu F N. An overview of guaranteed-loan industry in China. 2019. Qianzhan Industry Research Institute [2019-03-13]. https://www.qianzhan.com/analyst/detail/220/190312-07aee546.html. Google Scholar

[4] Li W, Ben S, Hommel U. Default contagion and systemic risk in loan guarantee networks. Account Finance, 2019, 59: 1923-1946 CrossRef Google Scholar

[5] Khandani A E, Kim A J, Lo A W. Consumer credit-risk models via machine-learning algorithms. J Banking Finance, 2010, 34: 2767-2787 CrossRef Google Scholar

[6] Velickovic P, Cucurull G, Casanova A, et al. Graph attention networks. In: Proceedings of the 5th International Conference on Learning Representations, Vancouver, 2018. 1--12. Google Scholar

[7] Cheng D W, Tu Y, Ma Z W, et al. Risk assessment for networked-guarantee loans using high-order graph attention representation. In: Proceedings of the 28th International Joint Conference on Artificial Intelligence, Macao, 2019. 5822--5828. Google Scholar

[8] Cui P, Wang X, Pei J. A Survey on Network Embedding. IEEE Trans Knowl Data Eng, 2019, 31: 833-852 CrossRef Google Scholar

[9] Perozzi B, Al-Rfou R, Skiena S. Deepwalk: online learning of social representations. In: Proceedings of the 20th ACM SIGKDD international conference on Knowledge discovery and data mining, New York, 2014. 701--710. Google Scholar

[10] Tang J, Qu M, Wang M Z, et al. Line: Large-scale information network embedding. In: Proceedings of the 24th international conference on world wide web, Florence, 2015. 1067--1077. Google Scholar

[11] Grover A, Leskovec J. node2vec: scalable feature learning for networks. In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, 2016. 855--864. Google Scholar

[12] Wang D X, Cui P, Zhu W W. Structural deep network embedding. In: Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, San Francisco, 2016. 1225--1234. Google Scholar

[13] Liu J, He Z C, Wei L, et al. Content to node: self-translation network embedding. IEEE Transactions on Knowledge and Data Engineering, 2019, 1:1-14. https://doi.org/10.1145/3219819.3219988. Google Scholar

[14] Wang X, Zhang Y D, Shi C. Hyperbolic heterogeneous information network embedding. In: Proceedings of the AAAI Conference on Artificial Intelligence, Hawaii, 2019. 5337--5344. Google Scholar

[15] Defferrard M, Bresson X, Vandergheynst P. Convolutional neural networks on graphs with fast localized spectral filtering. In: Proceedings of Advances in Neural Information Processing Systems, Barcelona, 2016. 3844--3852. Google Scholar

[16] Geng X, Li Y G, Wang L Y, et al. Spatiotemporal multi-graph convolution network for ride-hailing demand forecasting. In: Proceedings of the AAAI Conference on Artificial Intelligence, Hawaii, 2019. 3656--3663. Google Scholar

[17] Cheng D W, Xiang S, Shang C C, et al. Spatio-temporal attention-based neural network for credit card fraud detection. In: Proceedings of the AAAI Conference on Artificial Intelligence, New York, 2020. 362--369. Google Scholar

[18] Cheng D W, Tu Y, Ma Z W. BHONEM: Binary High-Order Network Embedding Methods for Networked-Guarantee Loans. J Comput Sci Technol, 2019, 34: 657-669 CrossRef Google Scholar

[19] Detering N, Meyer-Brandis T, Panagiotou K. Managing Default Contagion in Inhomogeneous Financial Networks. SIAM J Finan Math, 2019, 10: 578-614 CrossRef Google Scholar

[20] Battiston S, Puliga M, Kaushik R. DebtRank: Too Central to Fail? Financial Networks, the FED and Systemic Risk. Sci Rep, 2012, 2: 541 CrossRef ADS Google Scholar

[21] Acemoglu D, Ozdaglar A, Tahbaz-Salehi A. Systemic Risk and Stability in Financial Networks. Am Economic Rev, 2015, 105: 564-608 CrossRef Google Scholar

[22] Bougheas S, Kirman A. Complex financial networks and systemic risk: A review. In: Complexity and Geographical Economics. Cham: Springer, 2015. 19: 115--139. Google Scholar

[23] Catanzaro M, Buchanan M. Network opportunity. Nat Phys, 2013, 9: 121-123 CrossRef ADS Google Scholar

[24] Meng X, Liu X H, Credit risk evaluation for loan guarantee chain in China. In: Proceedings of the 13th Credit Scoring and Credit Control Conference, Edinburgh, 2015. 1--42. Google Scholar

[25] Brohée S, Faust K, Lima-Mendez G. Network Analysis Tools: from biological networks to clusters and pathways. Nat Protoc, 2008, 3: 1616-1629 CrossRef Google Scholar

[26] Vaswani A, Shazeer N, Parmar N, et al. Attention is all you need. In: Proceedings of Advances in Neural Information Processing Systems, Long Beach, 2017. 5998--6008. Google Scholar

[27] Chung J, Gulcehre C, Cho K, et al. Gated feedback recurrent neural networks. In: Proceedings of International Conference on Machine Learning, Lille, 2015. 2067--2075. Google Scholar

[28] Chen X Y, Liu S J, Sun R Y, et al. On the convergence of a class of Adam-type algorithms for non-convex optimization. In: Proceedings of the 7th International Conference on Learning Representations, New Orleans, 2019. 1--30. Google Scholar

[29] Tan F, Hou X, Zhang J. A Deep Learning Approach to Competing Risks Representation in Peer-to-Peer Lending. IEEE Trans Neural Netw Learning Syst, 2019, 30: 1565-1574 CrossRef Google Scholar

[30] Wu F L, Zhang T Y, Souza A H, et al. Simplifying graph convolutional networks. In: Proceedings of International Conference on Machine Learning, Long Beach, 2019. 6861--6871. Google Scholar

[31] 2-H. Google Scholar

  • Figure 1

    Illustration of a typical guarantee network and the diffusion process. (a) A typical guarantee network; protectłinebreak (b) individual risk; (c) contagious risk; (d) a typical contagion path; (e) all contagion paths; (f) guarantee relationship table

  • Figure 2

    (Color online) The model framework of this paper proposed contagion path risk rating model based on deep attention neural network (CRDAN). (a) Multi-path attention network; (b) behavior attention network; (c) risk rating network

  • Figure 3

    The description of networked-guarantee loan datasets for contagion path risk rating

  • Figure 4

    (Color online) The Precision@k of each risk category ((a) low, (b) medium, (c) high and (d) loss) in new constructed contagion paths

  • Figure 5

    (Color online) Visualization of attention weights in four different views. (a) Attention weights on each month; (b) the number of inner-chain nodes; (c) the number of chain duration months; (d) the number of neighbor chains

  • Figure 6

    (Color online) The visualization of the real-world loan network and the predicted high-risk contagion paths. protectłinebreak (a) A real-world complex loan network; (b) contagion paths; (c) the Sankey diragram view of contagion paths

  • Table 1   Comparison results of different methods in contagion path risk rating
    Methods Normal Low Medium High Loss Macro-F1 Micro-F1 N features
    LR 0.7561 0.6099 0.4812 0.9605 0.7537 0.7354 0.7048 $\sim$500
    GBDT 0.7722 0.6422 0.5412 0.9635 0.8086 0.7626 0.7270 $\sim$500
    DNN 0.7820 0.6637 0.5598 0.9659 0.8081 0.7707 0.7394 $\sim$500
    LSTM 0.7817 0.6684 0.5704 0.9626 0.8046 0.7711 0.7412 $\sim$500
    SGC 0.7971 0.6916 0.6087 0.9679 0.8241 0.7895 0.7599 17
    GAT 0.8076 0.7164 0.6419 0.9673 0.8179 0.7952 0.7726 17
    HGAR 0.8129 0.7216 0.6522 0.9688 0.8393$^{**}$ 0.8067 0.7805 17
    CRDAN-noPA 0.8125 0.7224 0.6488 0.9690 0.8359 0.8050 0.7798 17
    CRDAN-noBA 0.8205 0.7323 0.6565 0.9673 0.8391 0.8083 0.7859 17
    CRDAN-all 0.8344$^{**}$ 0.7539$^{**}$ 0.6953$^{**}$ 0.9697$^{**}$ 0.8298 0.8189$^{**}$ 0.8023$^{**}$ 17
qqqq

Contact and support