{"status":"ok","message-type":"work","message-version":"1.0.0","message":{"indexed":{"date-parts":[[2026,8,25]],"date-time":"2026-08-25T08:37:13Z","timestamp":1787647033002,"version":"build-2736575974"},"reference-count":46,"publisher":"PeerJ","license":[{"start":{"date-parts":[[2026,4,3]],"date-time":"2026-04-03T00:00:00Z","timestamp":1775174400000},"content-version":"unspecified","delay-in-days":0,"URL":"https:\/\/creativecommons.org\/licenses\/by\/4.0\/"}],"content-domain":{"domain":[],"crossmark-restriction":false},"short-container-title":[],"abstract":"<jats:p>Effective financial risk management is essential for safeguarding individual institutions and maintaining systemic stability, as risks such as accumulation, contagion, and external shocks can rapidly propagate through the financial ecosystem. Predicting these risks requires models capable of integrating heterogeneous, multi-source data that characterize modern financial environments. This study presents a deep learning framework for financial risk prediction that leverages diverse features from the FinBench dataset, a comprehensive benchmark consolidating hundreds of financial datasets. The proposed framework adopts a multi-source pretraining and fine-tuning strategy to learn transferable, high-quality representations across varied financial tabular data. By integrating tabular, temporal, and categorical information through transformer-based architectures, the model captures complex cross-feature dependencies that traditional approaches often overlook. Extensive experiments show that our framework consistently outperforms classical machine learning methods and state-of-the-art tabular deep learning models across credit default, fraud detection, and customer churn prediction tasks. Ablation studies further demonstrate the critical roles of backbone architecture and pretraining design in achieving superior predictive performance. These findings highlight the potential of deep representation learning on multi-source financial data as a modern and robust solution for accurate, generalizable financial risk assessment.<\/jats:p>","DOI":"10.7717\/peerj-cs.3694","type":"journal-article","created":{"date-parts":[[2026,4,3]],"date-time":"2026-04-03T08:18:28Z","timestamp":1775204308000},"page":"e3694","source":"Crossref","is-referenced-by-count":1,"title":["Predicting financial risk using deep learning on multi-source data: a modern approach to financial management"],"prefix":"10.7717","volume":"12","author":[{"given":"Hua","family":"Wang","sequence":"first","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]},{"given":"Xinyang","family":"Zhang","sequence":"additional","affiliation":[],"role":[{"vocabulary":"crossref","role":"author"}]}],"member":"4443","published-online":{"date-parts":[[2026,4,3]]},"reference":[{"key":"10.7717\/peerj-cs.3694\/ref-1","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1803.08375","article-title":"Deep learning using rectified linear units (ReLU)","author":"Agarap","year":"2018"},{"issue":"7","key":"10.7717\/peerj-cs.3694\/ref-2","doi-asserted-by":"publisher","first-page":"295","DOI":"10.3390\/jrfm17070295","article-title":"Understanding regulatory changes: deep learning in sustainable finance and banking","volume":"17","author":"Anghel","year":"2024","journal-title":"Journal of Risk and Financial Management"},{"issue":"8","key":"10.7717\/peerj-cs.3694\/ref-3","doi-asserted-by":"publisher","first-page":"6679","DOI":"10.1609\/aaai.v35i8.16826","article-title":"TabNet: attentive interpretable tabular learning","volume":"35","author":"Arik","year":"2021","journal-title":"Proceedings of the AAAI Conference on Artificial Intelligence"},{"key":"10.7717\/peerj-cs.3694\/ref-4","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1607.06450","article-title":"Layer normalization","author":"Ba","year":"2016"},{"issue":"2","key":"10.7717\/peerj-cs.3694\/ref-5","doi-asserted-by":"publisher","first-page":"423","DOI":"10.1109\/tpami.2018.2798607","article-title":"Multimodal machine learning: a survey and taxonomy","volume":"41","author":"Baltrusaitis","year":"2019","journal-title":"IEEE Transactions on Pattern Analysis and Machine Intelligence"},{"issue":"6","key":"10.7717\/peerj-cs.3694\/ref-6","doi-asserted-by":"publisher","first-page":"7499","DOI":"10.1109\/tnnls.2022.3229161","article-title":"Deep neural networks and tabular data: a survey","volume":"35","author":"Borisov","year":"2024","journal-title":"IEEE Transactions on Neural Networks and Learning Systems"},{"issue":"1","key":"10.7717\/peerj-cs.3694\/ref-7","doi-asserted-by":"publisher","first-page":"5","DOI":"10.1007\/978-3-030-56485-8_3","article-title":"Random forests","volume":"45","author":"Breiman","year":"2001","journal-title":"Machine learning"},{"key":"10.7717\/peerj-cs.3694\/ref-8","doi-asserted-by":"publisher","DOI":"10.1145\/2939672.2939785","article-title":"XGBoost: a scalable tree boosting system","author":"Chen","year":"2016"},{"key":"10.7717\/peerj-cs.3694\/ref-9","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2012.15330","article-title":"Sequential deep learning for credit risk monitoring with tabular financial data","author":"Clements","year":"2020"},{"key":"10.7717\/peerj-cs.3694\/ref-10","doi-asserted-by":"publisher","first-page":"50426","DOI":"10.1109\/access.2021.3068854","article-title":"Making deep learning-based predictions for credit scoring explainable","volume":"9","author":"Dastile","year":"2021","journal-title":"IEEE Access"},{"key":"10.7717\/peerj-cs.3694\/ref-11","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1810.11363","article-title":"CatBoost: gradient boosting with categorical features support","author":"Dorogush","year":"2018"},{"issue":"2","key":"10.7717\/peerj-cs.3694\/ref-12","doi-asserted-by":"publisher","first-page":"215","DOI":"10.35741\/issn.0258-2724.59.2.14","article-title":"Models for financial crisis prediction using deep learning techniques","volume":"59","author":"Elnaggar","year":"2024","journal-title":"Journal of Southwest Jiaotong University"},{"key":"10.7717\/peerj-cs.3694\/ref-13","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2106.11959","article-title":"Revisiting deep learning models for tabular data","author":"Gorishniy","year":"2021"},{"key":"10.7717\/peerj-cs.3694\/ref-14","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1703.04247","article-title":"DeepFM: a factorization-machine based neural network for CTR prediction","author":"Guo","year":"2017"},{"key":"10.7717\/peerj-cs.3694\/ref-15","doi-asserted-by":"publisher","first-page":"63373","DOI":"10.1109\/access.2019.2916887","article-title":"Deep multimodal representation learning: a survey","volume":"7","author":"Guo","year":"2019","journal-title":"IEEE Access"},{"issue":"19","key":"10.7717\/peerj-cs.3694\/ref-16","doi-asserted-by":"publisher","first-page":"3181","DOI":"10.3390\/electronics11193181","article-title":"Emerging trends in deep learning for credit scoring: a review","volume":"11","author":"Hayashi","year":"2022","journal-title":"Electronics"},{"key":"10.7717\/peerj-cs.3694\/ref-17","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2012.06678","article-title":"TabTransformer: tabular data modeling using contextual embeddings","author":"Huang","year":"2020"},{"issue":"5","key":"10.7717\/peerj-cs.3694\/ref-18","doi-asserted-by":"publisher","first-page":"5601","DOI":"10.11591\/ijece.v14i5.pp5601-5611","article-title":"Forecasting stock market prices using deep learning methods","volume":"14","author":"Ismailova","year":"2024","journal-title":"International Journal of Electrical and Computer Engineering (IJECE)"},{"key":"10.7717\/peerj-cs.3694\/ref-19","first-page":"3149","article-title":"LightGBM: a highly efficient gradient boosting decision tree","author":"Ke","year":"2017"},{"key":"10.7717\/peerj-cs.3694\/ref-20","first-page":"384","article-title":"DeepGBM: a deep learning framework distilled by GBDT for online prediction tasks","author":"Ke","year":"2019"},{"issue":"2","key":"10.7717\/peerj-cs.3694\/ref-21","doi-asserted-by":"publisher","first-page":"115060","DOI":"10.1016\/j.eswa.2021.115060","article-title":"Robo-advisor using genetic algorithm and BERT sentiments from tweets for hybrid portfolio optimisation","volume":"179","author":"Leow","year":"2021","journal-title":"Expert Systems with Applications"},{"issue":"8","key":"10.7717\/peerj-cs.3694\/ref-22","doi-asserted-by":"publisher","first-page":"e0328013","DOI":"10.1371\/journal.pone.0328013","article-title":"LTR-Net: a deep learning-based approach for financial data prediction and risk evaluation in enterprises","volume":"20","author":"Liu","year":"2025","journal-title":"PLOS ONE"},{"key":"10.7717\/peerj-cs.3694\/ref-23","doi-asserted-by":"publisher","first-page":"1","DOI":"10.1109\/tim.2025.3562973","article-title":"Establishing a digital twin diagnostic model based on cross-device transfer learning","volume":"74","author":"Liu","year":"2025","journal-title":"IEEE Transactions on Instrumentation and Measurement"},{"issue":"14","key":"10.7717\/peerj-cs.3694\/ref-24","doi-asserted-by":"publisher","first-page":"5232","DOI":"10.3390\/en15145232","article-title":"Application of temporal fusion transformer for day-ahead PV power forecasting","volume":"15","author":"L\u00f3pez Santos","year":"2022","journal-title":"Energies"},{"issue":"1","key":"10.7717\/peerj-cs.3694\/ref-25","doi-asserted-by":"publisher","first-page":"82","DOI":"10.3126\/jes2.v2i1.60399","article-title":"Multi-class credit risk analysis using deep learning","volume":"2","author":"Paudel","year":"2023","journal-title":"Journal of Engineering and Sciences"},{"issue":"6","key":"10.7717\/peerj-cs.3694\/ref-26","doi-asserted-by":"publisher","first-page":"8","DOI":"10.55640\/business\/volume06issue06-02","article-title":"Forecasting bank profitability using deep learning and macroeconomic indicators: a comparative model study","volume":"6","author":"Sajal","year":"2025","journal-title":"International Interdisciplinary Business Economics Advancement Journal"},{"issue":"3","key":"10.7717\/peerj-cs.3694\/ref-27","doi-asserted-by":"publisher","first-page":"85","DOI":"10.1016\/j.neunet.2014.09.003","article-title":"Deep learning in neural networks: an overview","volume":"61","author":"Schmidhuber","year":"2015","journal-title":"Neural Networks"},{"issue":"1","key":"10.7717\/peerj-cs.3694\/ref-28","doi-asserted-by":"publisher","first-page":"111","DOI":"10.56696\/ijamer.v2i1.30","article-title":"Optimizing banking stock price prediction: deep learning based approach","volume":"2","author":"Setiawan","year":"2024","journal-title":"International Journal of Accounting, Management, and Economics Research"},{"issue":"3","key":"10.7717\/peerj-cs.3694\/ref-29","doi-asserted-by":"publisher","first-page":"1712","DOI":"10.29121\/shodhkosh.v5.i3.2024.5185","article-title":"Enhancing market trend forecasting with explainable AI: a comparative analysis of deep learning models and interpretability techniques","volume":"5","author":"Sisodia","year":"2024","journal-title":"ShodhKosh: Journal of Visual and Performing Arts"},{"key":"10.7717\/peerj-cs.3694\/ref-30","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2106.01342","article-title":"SAINT: improved neural networks for tabular data via row attention and contrastive pre-training","author":"Somepalli","year":"2021"},{"key":"10.7717\/peerj-cs.3694\/ref-31","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1810.11921","article-title":"AutoInt: automatic feature interaction learning via self-attentive neural networks","author":"Song","year":"2018"},{"issue":"11","key":"10.7717\/peerj-cs.3694\/ref-32","doi-asserted-by":"publisher","first-page":"16687","DOI":"10.1109\/tits.2024.3409874","article-title":"Subjective driving risk prediction based on spatiotemporal distribution features of human driver\u2019s cognitive risk","volume":"25","author":"Song","year":"2024","journal-title":"IEEE Transactions on Intelligent Transportation Systems"},{"issue":"1","key":"10.7717\/peerj-cs.3694\/ref-33","doi-asserted-by":"publisher","first-page":"78","DOI":"10.35746\/jtim.v7i1.650","article-title":"Comparative analysis of stock price prediction using deep learning with data scaling method","volume":"7","author":"Switrayana","year":"2025","journal-title":"JTIM: Jurnal Teknologi Informasi dan Multimedia"},{"key":"10.7717\/peerj-cs.3694\/ref-34","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.1706.03762","article-title":"Attention is all you need","author":"Vaswani","year":"2017"},{"key":"10.7717\/peerj-cs.3694\/ref-35","doi-asserted-by":"publisher","first-page":"284","DOI":"10.1016\/j.jmsy.2025.06.018","article-title":"A mutual cross-attention fusion network for surface roughness prediction in robotic machining process using internal and external signals","volume":"82","author":"Wang","year":"2025","journal-title":"Journal of Manufacturing Systems"},{"issue":"14","key":"10.7717\/peerj-cs.3694\/ref-36","doi-asserted-by":"publisher","first-page":"5964","DOI":"10.3390\/su16145964","article-title":"Application of a multi-model fusion forecasting approach in runoff prediction: a case study of the Yangtze river source region","volume":"16","author":"Wang","year":"2024","journal-title":"Sustainability"},{"issue":"1","key":"10.7717\/peerj-cs.3694\/ref-37","doi-asserted-by":"publisher","first-page":"107208","DOI":"10.1016\/j.neunet.2025.107208","article-title":"Deterministic autoencoder using Wasserstein loss for tabular data generation","volume":"185","author":"Wang","year":"2025a","journal-title":"Neural Networks"},{"issue":"6","key":"10.7717\/peerj-cs.3694\/ref-38","doi-asserted-by":"publisher","first-page":"104292","DOI":"10.1016\/j.artint.2025.104292","article-title":"TTVAE: transformer-based generative modeling for tabular data generation","volume":"340","author":"Wang","year":"2025b","journal-title":"Artificial Intelligence"},{"key":"10.7717\/peerj-cs.3694\/ref-39","doi-asserted-by":"publisher","first-page":"46335","DOI":"10.1109\/access.2020.2974101","article-title":"Feature extraction and analysis of natural language processing for deep learning english language","volume":"8","author":"Wang","year":"2020","journal-title":"IEEE Access"},{"issue":"1","key":"10.7717\/peerj-cs.3694\/ref-40","doi-asserted-by":"publisher","first-page":"1","DOI":"10.2478\/amns-2025-0818","article-title":"Emotional analysis and semantic understanding of multimodal network language data","volume":"10","author":"Weimiao","year":"2025","journal-title":"Applied Mathematics and Nonlinear Sciences"},{"key":"10.7717\/peerj-cs.3694\/ref-41","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2108.09084","article-title":"Fastformer: additive attention can be all you need","author":"Wu","year":"2021"},{"issue":"5","key":"10.7717\/peerj-cs.3694\/ref-42","doi-asserted-by":"publisher","first-page":"1","DOI":"10.55041\/ijsrem48416","article-title":"LSTM-RNN based predictive stock analysis with real-time sentiment interpretation","volume":"9","author":"Yadhunandan","year":"2025","journal-title":"International Journal of Scientific Research in Engineering and Management"},{"key":"10.7717\/peerj-cs.3694\/ref-43","doi-asserted-by":"publisher","DOI":"10.48550\/arXiv.2308.00065","article-title":"FinPT: Financial risk prediction with profile tuning on pretrained foundation models","author":"Yin","year":"2023"},{"key":"10.7717\/peerj-cs.3694\/ref-44","first-page":"11033","article-title":"VIME: extending the success of self- and semi-supervised learning to tabular domain","volume":"33","author":"Yoon","year":"2020"},{"issue":"2","key":"10.7717\/peerj-cs.3694\/ref-45","doi-asserted-by":"publisher","first-page":"1","DOI":"10.54097\/c6jmgg90","article-title":"Research on the application of multimodal information fusion in financial data prediction","volume":"4","author":"Zhi","year":"2025","journal-title":"Journal of Computer Science and Artificial Intelligence"},{"key":"10.7717\/peerj-cs.3694\/ref-46","doi-asserted-by":"publisher","DOI":"10.21203\/rs.3.rs-7521327\/v1","article-title":"Deep temporal features and multi-level cross-modal attention fusion for multimodal sentiment analysis","author":"Zhu","year":"2025"}],"container-title":["PeerJ Computer Science"],"original-title":[],"language":"en","link":[{"URL":"https:\/\/peerj.com\/articles\/cs-3694.pdf","content-type":"application\/pdf","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/peerj.com\/articles\/cs-3694.xml","content-type":"application\/xml","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/peerj.com\/articles\/cs-3694.html","content-type":"text\/html","content-version":"vor","intended-application":"text-mining"},{"URL":"https:\/\/peerj.com\/articles\/cs-3694.pdf","content-type":"unspecified","content-version":"vor","intended-application":"similarity-checking"}],"deposited":{"date-parts":[[2026,4,3]],"date-time":"2026-04-03T08:18:33Z","timestamp":1775204313000},"score":1,"resource":{"primary":{"URL":"https:\/\/peerj.com\/articles\/cs-3694"}},"subtitle":[],"short-title":[],"issued":{"date-parts":[[2026,4,3]]},"references-count":46,"alternative-id":["10.7717\/peerj-cs.3694"],"URL":"https:\/\/doi.org\/10.7717\/peerj-cs.3694","archive":["CLOCKSS","LOCKSS","Portico"],"relation":{},"ISSN":["2376-5992"],"issn-type":[{"value":"2376-5992","type":"electronic"}],"subject":[],"published":{"date-parts":[[2026,4,3]]},"article-number":"e3694"}}