Deep Dense Convolutional Networks for Repayment Prediction in Peer-to-Peer Lending

Ji Yoon Kim, Sung-Bae Cho

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Abstract

In peer-to-peer (P2P) lending, it is important to predict default of borrowers because the lenders would suffer financial loss if the borrower fails to pay money. The huge lending transaction data generated online helps to predict repayment of the borrowers, but there are limitations in extracting features based on the complex information. Convolutional neural networks (CNN) can automatically extract useful features from large P2P lending data. However, as deep CNN becomes more complex and deeper, the information about input vanishes and overfitting occurs. In this paper, we propose a deep dense convolutional networks (DenseNet) for default prediction in P2P social lending to automatically extract features and improve the performance. DenseNet ensures the flow of loan information through dense connectivity and automatically extracts discriminative features with convolution and pooling operations. We capture the complex features of lending data and reuse loan information to predict the repayment of the borrower. Experimental results show that the proposed method automatically extracts useful features from Lending Club data, avoids overfitting, and is effective in default prediction. In comparison with deep CNN and other machine learning methods, the proposed method has achieved the highest performance with 79.6%. We demonstrate the usefulness of the proposed method as the 5-fold cross-validation to evaluate the performance.

Original languageEnglish
Title of host publicationInternational Joint Conference SOCO’18-CISIS’18-ICEUTE’18, Proceedings
EditorsJose Antonio Saez, Emilio Corchado, Alvaro Herrero, Manuel Grana, Jose Manuel Lopez-Guede, Oier Etxaniz, Hector Quintian
PublisherSpringer Verlag
Pages134-144
Number of pages11
ISBN (Print)9783319941196
DOIs
Publication statusPublished - 2019 Jan 1
EventInternational Joint Conference: 13th International Conference on Soft Computing Models, SOCO 2018, 11th International Conference on Computational Intelligence in Security for Information Systems, CISIS 2018 and 9th International Conference on EUropean Transnational Education, ICEUTE 2018 - san sebastian, Spain
Duration: 2018 Jun 62018 Jun 8

Publication series

NameAdvances in Intelligent Systems and Computing
Volume771
ISSN (Print)2194-5357

Other

OtherInternational Joint Conference: 13th International Conference on Soft Computing Models, SOCO 2018, 11th International Conference on Computational Intelligence in Security for Information Systems, CISIS 2018 and 9th International Conference on EUropean Transnational Education, ICEUTE 2018
CountrySpain
Citysan sebastian
Period18/6/618/6/8

Fingerprint

Neural networks
Convolution
Learning systems

All Science Journal Classification (ASJC) codes

  • Control and Systems Engineering
  • Computer Science(all)

Cite this

Kim, J. Y., & Cho, S-B. (2019). Deep Dense Convolutional Networks for Repayment Prediction in Peer-to-Peer Lending. In J. A. Saez, E. Corchado, A. Herrero, M. Grana, J. M. Lopez-Guede, O. Etxaniz, & H. Quintian (Eds.), International Joint Conference SOCO’18-CISIS’18-ICEUTE’18, Proceedings (pp. 134-144). (Advances in Intelligent Systems and Computing; Vol. 771). Springer Verlag. https://doi.org/10.1007/978-3-319-94120-2_13
Kim, Ji Yoon ; Cho, Sung-Bae. / Deep Dense Convolutional Networks for Repayment Prediction in Peer-to-Peer Lending. International Joint Conference SOCO’18-CISIS’18-ICEUTE’18, Proceedings. editor / Jose Antonio Saez ; Emilio Corchado ; Alvaro Herrero ; Manuel Grana ; Jose Manuel Lopez-Guede ; Oier Etxaniz ; Hector Quintian. Springer Verlag, 2019. pp. 134-144 (Advances in Intelligent Systems and Computing).
@inproceedings{2dfebbfea1884747a35424ae4b8b85a4,
title = "Deep Dense Convolutional Networks for Repayment Prediction in Peer-to-Peer Lending",
abstract = "In peer-to-peer (P2P) lending, it is important to predict default of borrowers because the lenders would suffer financial loss if the borrower fails to pay money. The huge lending transaction data generated online helps to predict repayment of the borrowers, but there are limitations in extracting features based on the complex information. Convolutional neural networks (CNN) can automatically extract useful features from large P2P lending data. However, as deep CNN becomes more complex and deeper, the information about input vanishes and overfitting occurs. In this paper, we propose a deep dense convolutional networks (DenseNet) for default prediction in P2P social lending to automatically extract features and improve the performance. DenseNet ensures the flow of loan information through dense connectivity and automatically extracts discriminative features with convolution and pooling operations. We capture the complex features of lending data and reuse loan information to predict the repayment of the borrower. Experimental results show that the proposed method automatically extracts useful features from Lending Club data, avoids overfitting, and is effective in default prediction. In comparison with deep CNN and other machine learning methods, the proposed method has achieved the highest performance with 79.6{\%}. We demonstrate the usefulness of the proposed method as the 5-fold cross-validation to evaluate the performance.",
author = "Kim, {Ji Yoon} and Sung-Bae Cho",
year = "2019",
month = "1",
day = "1",
doi = "10.1007/978-3-319-94120-2_13",
language = "English",
isbn = "9783319941196",
series = "Advances in Intelligent Systems and Computing",
publisher = "Springer Verlag",
pages = "134--144",
editor = "Saez, {Jose Antonio} and Emilio Corchado and Alvaro Herrero and Manuel Grana and Lopez-Guede, {Jose Manuel} and Oier Etxaniz and Hector Quintian",
booktitle = "International Joint Conference SOCO’18-CISIS’18-ICEUTE’18, Proceedings",
address = "Germany",

}

Kim, JY & Cho, S-B 2019, Deep Dense Convolutional Networks for Repayment Prediction in Peer-to-Peer Lending. in JA Saez, E Corchado, A Herrero, M Grana, JM Lopez-Guede, O Etxaniz & H Quintian (eds), International Joint Conference SOCO’18-CISIS’18-ICEUTE’18, Proceedings. Advances in Intelligent Systems and Computing, vol. 771, Springer Verlag, pp. 134-144, International Joint Conference: 13th International Conference on Soft Computing Models, SOCO 2018, 11th International Conference on Computational Intelligence in Security for Information Systems, CISIS 2018 and 9th International Conference on EUropean Transnational Education, ICEUTE 2018, san sebastian, Spain, 18/6/6. https://doi.org/10.1007/978-3-319-94120-2_13

Deep Dense Convolutional Networks for Repayment Prediction in Peer-to-Peer Lending. / Kim, Ji Yoon; Cho, Sung-Bae.

International Joint Conference SOCO’18-CISIS’18-ICEUTE’18, Proceedings. ed. / Jose Antonio Saez; Emilio Corchado; Alvaro Herrero; Manuel Grana; Jose Manuel Lopez-Guede; Oier Etxaniz; Hector Quintian. Springer Verlag, 2019. p. 134-144 (Advances in Intelligent Systems and Computing; Vol. 771).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

TY - GEN

T1 - Deep Dense Convolutional Networks for Repayment Prediction in Peer-to-Peer Lending

AU - Kim, Ji Yoon

AU - Cho, Sung-Bae

PY - 2019/1/1

Y1 - 2019/1/1

N2 - In peer-to-peer (P2P) lending, it is important to predict default of borrowers because the lenders would suffer financial loss if the borrower fails to pay money. The huge lending transaction data generated online helps to predict repayment of the borrowers, but there are limitations in extracting features based on the complex information. Convolutional neural networks (CNN) can automatically extract useful features from large P2P lending data. However, as deep CNN becomes more complex and deeper, the information about input vanishes and overfitting occurs. In this paper, we propose a deep dense convolutional networks (DenseNet) for default prediction in P2P social lending to automatically extract features and improve the performance. DenseNet ensures the flow of loan information through dense connectivity and automatically extracts discriminative features with convolution and pooling operations. We capture the complex features of lending data and reuse loan information to predict the repayment of the borrower. Experimental results show that the proposed method automatically extracts useful features from Lending Club data, avoids overfitting, and is effective in default prediction. In comparison with deep CNN and other machine learning methods, the proposed method has achieved the highest performance with 79.6%. We demonstrate the usefulness of the proposed method as the 5-fold cross-validation to evaluate the performance.

AB - In peer-to-peer (P2P) lending, it is important to predict default of borrowers because the lenders would suffer financial loss if the borrower fails to pay money. The huge lending transaction data generated online helps to predict repayment of the borrowers, but there are limitations in extracting features based on the complex information. Convolutional neural networks (CNN) can automatically extract useful features from large P2P lending data. However, as deep CNN becomes more complex and deeper, the information about input vanishes and overfitting occurs. In this paper, we propose a deep dense convolutional networks (DenseNet) for default prediction in P2P social lending to automatically extract features and improve the performance. DenseNet ensures the flow of loan information through dense connectivity and automatically extracts discriminative features with convolution and pooling operations. We capture the complex features of lending data and reuse loan information to predict the repayment of the borrower. Experimental results show that the proposed method automatically extracts useful features from Lending Club data, avoids overfitting, and is effective in default prediction. In comparison with deep CNN and other machine learning methods, the proposed method has achieved the highest performance with 79.6%. We demonstrate the usefulness of the proposed method as the 5-fold cross-validation to evaluate the performance.

UR - http://www.scopus.com/inward/record.url?scp=85048600530&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85048600530&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-94120-2_13

DO - 10.1007/978-3-319-94120-2_13

M3 - Conference contribution

SN - 9783319941196

T3 - Advances in Intelligent Systems and Computing

SP - 134

EP - 144

BT - International Joint Conference SOCO’18-CISIS’18-ICEUTE’18, Proceedings

A2 - Saez, Jose Antonio

A2 - Corchado, Emilio

A2 - Herrero, Alvaro

A2 - Grana, Manuel

A2 - Lopez-Guede, Jose Manuel

A2 - Etxaniz, Oier

A2 - Quintian, Hector

PB - Springer Verlag

ER -

Kim JY, Cho S-B. Deep Dense Convolutional Networks for Repayment Prediction in Peer-to-Peer Lending. In Saez JA, Corchado E, Herrero A, Grana M, Lopez-Guede JM, Etxaniz O, Quintian H, editors, International Joint Conference SOCO’18-CISIS’18-ICEUTE’18, Proceedings. Springer Verlag. 2019. p. 134-144. (Advances in Intelligent Systems and Computing). https://doi.org/10.1007/978-3-319-94120-2_13