Research on Transformation and Interpretability in Credit Classification
1
조회
0
다운로드
0
인용
0
참고문헌
키워드
초록
The modern financial industry demands rapid decision-making based on diverse information from dynamicenvironments. Predicting outcomes from such data is complex due to rapid shifts influenced by numerousfactors. Despite advancements in artificial intelligence technology that offer sophisticated analytical models,accurately predicting outcomes and providing sufficient justification for these predictions remain challenging,particularly with fragmented model constructions. In this paper, we propose a novel approach for efficientprocessing of available public personal credit data, deriving new analysis elements, and comparing predictioninterpretations. Specifically, we develop 11 prediction models that can be categorized into two types: dataimage transformation and time-series transformation. The models undergo standardization, preprocessing, andcross-validation for optimization, with their predictive performances compared and validated. Modelsleveraging convolutional neural network (CNN) and convolutional neural network-long short-term memory(CNN-LSTM) architectures demonstrate strong performance across both categories. To fully interpret theclassification process, SHAP is applied to compare and explain the prediction results for each model type.
Abstract
The modern financial industry demands rapid decision-making based on diverse information from dynamicenvironments. Predicting outcomes from such data is complex due to rapid shifts influenced by numerousfactors. Despite advancements in artificial intelligence technology that offer sophisticated analytical models,accurately predicting outcomes and providing sufficient justification for these predictions remain challenging,particularly with fragmented model constructions. In this paper, we propose a novel approach for efficientprocessing of available public personal credit data, deriving new analysis elements, and comparing predictioninterpretations. Specifically, we develop 11 prediction models that can be categorized into two types: dataimage transformation and time-series transformation. The models undergo standardization, preprocessing, andcross-validation for optimization, with their predictive performances compared and validated. Modelsleveraging convolutional neural network (CNN) and convolutional neural network-long short-term memory(CNN-LSTM) architectures demonstrate strong performance across both categories. To fully interpret theclassification process, SHAP is applied to compare and explain the prediction results for each model type.
참고문헌 (31)
- A. Bachmann, "Online peer-to-peer lending : a literature review", Journal of Internet Banking and Commerce, 16(2), pp.1~18 (2011)
- G. G. Parker, "Platform Revolution: How Networked Markets Are Transforming the Economy and How to Make Them Work for You" (2017), W. W. Norton & Company
- S. Sohangir, "Big data : deep learning for financial sentiment analysis", Journal of Big Data, 5(1) (2018)[DOI]
- J. Y. Kim, "Towards repayment prediction in peer-to-peer social lending using deep learning", Mathematics, 7(11) (2019)[DOI]
- W. Jiao, "The image classification method with CNN-XGBoost model based on adaptive particle swarm optimization", Information, 12(4) (2021)[DOI]
- J. M. T. Wu, "A graph-based CNN-LSTM stock price prediction algorithm with leading indicators", Multimedia Systems, 29(3), pp.1751~1770 (2023)[DOI]
- H. Qian, "Soft reordering one-dimensional convolutional neural network for credit scoring", Knowledge-Based Systems, 266 (2023)[DOI]
- C. Wang, "A deep learning approach for credit scoring of peer-to-peer lending using attention mechanism LSTM", IEEE Access, 7, pp.2161~2168 (2018)[DOI]
- L. Liang, "Forecasting peer-to-peer platform default rate with LSTM neural network", Electronic Commerce Research and Applications, 43 (2020)[DOI]
- T. Gramespacher, "Employing explainable AI to optimize the return target function of a loan portfolio", Frontiers in Artificial Intelligence, 4 (2021)[DOI]
- "Towards Explainable Artificial Intelligence in banking and financial services" (2021)[URL]
- I. D. C. Arifah, "Artificial intelligence in credit risk management of peer-to-peer lending financial technology : systematic literature review", Proceedings of 2023 6th International Conference of Computer and Informatics Engineering (IC2IE), pp.329~334 (2023)
- "CNNPred: CNN-based stock market prediction using several data sources" (2018)[URL]
- J. Cao, "Stock price forecasting model based on modified convolution neural network and financial time series analysis", International Journal of Communication Systems, 32(12) (2019)[DOI]
- H. S. Lee, "LSTM-based deep learning for time series forecasting : the case of corporate credit score prediction", The Journal of Information Systems, 29(1), pp.241~265 (2002)[DOI]
- S. Liu, "CNN-LSTM neural network model for quantitative strategy analysis in stock markets", Neural Information Processing, pp.198~206 (2017), Springer
- J. Eapen, "Novel deep learning model with CNN and bi-directional LSTM for improved stock market index prediction", Proceedings of 2019 IEEE 9th annual computing and communication workshop and conference (CCWC), pp.264~270 (2019)
- W. Lu, "A CNN-BiLSTM-AM method for stock price prediction", Neural Computing and Applications, 33(10), pp.4741~4753 (2021)[DOI]
- M. T. Ribeiro, "Why should I trust you? Explaining the predictions of any classifier", Proceedings of the 22nd ACM SIGKDD International Conference on Knowledge Discovery and Data Mining, pp.1135~1144 (2016)
- S. M. Lundberg, "A unified approach to interpreting model predictions", Advances in Neural Information Processing Systems, 30, pp.4765~4774 (2017)
- N. Bussmann, "Explainable AI in fintech risk management", Frontiers in Artificial Intelligence, 3 (2020)[DOI]
- "Explainable AI in credit risk management" (2021)[URL]
- A. Gramegna, "SHAP and LIME : an evaluation of discriminative power in credit risk", Frontiers in Artificial Intelligence, 4 (2021)[DOI]
- J. Kim, "A study on credit risk predicting base on XAI using transformation of time series data", Ph. D., Department of Convergence Technology, pp.2024
- "An interpretable model with globally consistent explanations for credit risk" (2018)[URL]
- Y. Zhu, "Converting tabular data into images for deep learning with convolutional neural networks", Scientific Reports, 11(1) (2021)[DOI]
- A. Sharma, "DeepInsight : a methodology to transform a non-image data to an image for convolution neural network architecture", Scientific Reports, 9(1) (2019)[DOI]
- A. Tharwat, "Classification assessment methods", Applied Computing and Informatics, 17(1), pp.168~192 (2021)[DOI]
- K. Potdar, "A comparative study of categorical variable encoding techniques for neural network classifiers", International Journal of Computer Applications, 175(4), pp.7~9 (2017)[DOI]
- X. Fu, "A sentiment-aware trading volume prediction model for P2P market using LSTM", IEEE Access, 7, pp.81934~81944 (2019)[DOI]
- J. Fan, "Predicting daily diffuse horizontal solar radiation in various climatic regions of China using support vector machine and tree-based soft computing models with local and extrinsic climatic data", Journal of Cleaner Production, 248 (2020)[DOI]