Sorry, you need to enable JavaScript to visit this website.

facebooktwittermailshare

Enhanced Recurrent Neural Network for Combining Static and Dynamic Features for Credit Card Default Prediction

Abstract: 

Deep learning models have been shown to be capable of extracting high-level representations from the increasing amount of customer-level data generated via fast-growing financial activities. In financial data, dynamic features that evolve with time are commonly observed. However, such time dependencies are often ignored in classical classification models. In this study, we propose to learn a Recurrent Neural Network (RNN) feature extractor with GRU on credit card payment history to leverage the time dependencies embedded in these dynamic features. Input sequences are first preprocessed by this feature extractor. The extracted dynamic features along with the static features are then utilized to train an enhanced RNN model (RNN-RF) to predict credit card client defaults. Numerical experiments confirmed that the enhanced RNN predictor indeed provides the best performance in both lift index (0.659) and AUC (0.782) compared to the other benchmark models. The proposed model allows us to effectively combine static and dynamic features to provide superior predictive performance for financial data.

up
0 users have voted:

Paper Details

Authors:
Te-Cheng Hsu, Shing-Tzuo Liou, Yun-Ping Wang, Yung-Shun Huang, Che Lin
Submitted On:
8 May 2019 - 9:27am
Short Link:
Type:
Poster
Event:
Presenter's Name:
Te-Cheng Hsu
Paper Code:
DISPS-P3
Document Year:
2019
Cite

Document Files

poster_new_0415.pdf

(53)

Subscribe

[1] Te-Cheng Hsu, Shing-Tzuo Liou, Yun-Ping Wang, Yung-Shun Huang, Che Lin, "Enhanced Recurrent Neural Network for Combining Static and Dynamic Features for Credit Card Default Prediction", IEEE SigPort, 2019. [Online]. Available: http://sigport.org/4093. Accessed: Dec. 07, 2019.
@article{4093-19,
url = {http://sigport.org/4093},
author = {Te-Cheng Hsu; Shing-Tzuo Liou; Yun-Ping Wang; Yung-Shun Huang; Che Lin },
publisher = {IEEE SigPort},
title = {Enhanced Recurrent Neural Network for Combining Static and Dynamic Features for Credit Card Default Prediction},
year = {2019} }
TY - EJOUR
T1 - Enhanced Recurrent Neural Network for Combining Static and Dynamic Features for Credit Card Default Prediction
AU - Te-Cheng Hsu; Shing-Tzuo Liou; Yun-Ping Wang; Yung-Shun Huang; Che Lin
PY - 2019
PB - IEEE SigPort
UR - http://sigport.org/4093
ER -
Te-Cheng Hsu, Shing-Tzuo Liou, Yun-Ping Wang, Yung-Shun Huang, Che Lin. (2019). Enhanced Recurrent Neural Network for Combining Static and Dynamic Features for Credit Card Default Prediction. IEEE SigPort. http://sigport.org/4093
Te-Cheng Hsu, Shing-Tzuo Liou, Yun-Ping Wang, Yung-Shun Huang, Che Lin, 2019. Enhanced Recurrent Neural Network for Combining Static and Dynamic Features for Credit Card Default Prediction. Available at: http://sigport.org/4093.
Te-Cheng Hsu, Shing-Tzuo Liou, Yun-Ping Wang, Yung-Shun Huang, Che Lin. (2019). "Enhanced Recurrent Neural Network for Combining Static and Dynamic Features for Credit Card Default Prediction." Web.
1. Te-Cheng Hsu, Shing-Tzuo Liou, Yun-Ping Wang, Yung-Shun Huang, Che Lin. Enhanced Recurrent Neural Network for Combining Static and Dynamic Features for Credit Card Default Prediction [Internet]. IEEE SigPort; 2019. Available from : http://sigport.org/4093