International Journal of Social Sciences and Economic Management, 2026, 7(1); doi: 10.3807/IJSSEM.2026.070112.
Wei Sun
School of Computer and Big Data, Jining Normal University, Ulanqab 012000, Inner Mongolia, China
In the context of financial information explosion, unstructured text (such as market news, corporate financial reports, etc.) contains key event information, and event extraction technology becomes the core bridge for quantitative analysis through structured "who, when, where, and what" elements. However, traditional methods face three major challenges: syntactic complexity, dynamic evolution of terminology, and intertwining of multiple event entities. Existing research has limitations such as inconsistent standards, ineffective cross sentence causal capture, and dependence on annotated data. This study proposes a document level financial event extraction method based on dependency syntax enhancement and BKP algorithm optimization: integrating syntactic dependency relationships and part of speech/entity type information through graph attention networks to enhance the semantic integrity of argument extraction; Using RoBERTa to capture document level context, combined with importance score optimization and dynamic pruning strategy, to achieve efficient combination decoding and reduce redundant computation. Empirical evidence on the publicly available CFA dataset shows that the new model significantly outperforms existing benchmarks in terms of accuracy (improved by 8% -12%) and processing speed (improved by 30% -50%), especially in multi event text scenarios. Research has found that the syntactic semantic enhancement module effectively solves the problem of extracting semantic integrity from complex causal sentences. The BKP algorithm reduces the complexity of multi event decoding through path optimization and dynamic pruning, while maintaining a high recall rate. This study achieved semantic coherence from sentence level to document level, providing a new paradigm for dynamic information extraction in financial scenarios. In the future, the model will be extended to non-financial fields for generalization verification, and visualization tools will be developed to enhance interpretability.
Financial document event extraction, dependency syntax enhancement; BKP algorithm optimization, combined decoding, document level language, seamless integration
Wei Sun. Research on Financial Document Event Extraction Method Based on Dependency Syntax Enhancement and BKP Algorithm Optimization. International Journal of Social Sciences and Economic Management (2026), Vol. 7, Issue 1: 105-114. https://doi.doi.org/10.3807/IJSSEM.2026.070112.
[1] Hui, X. (2026). Research on the Design and Optimization of Automated Data Collection and Visual Dashboard in the Medical Industry. Journal of Computer, Signal, and System Research, 3(1), 27-34.
[2] Wang, Y. (2026). Research on Optimization of Neuromuscular Rehabilitation Program Based on Physiological Assessment. European Journal of AI, Computing & Informatics, 2(1), 21-30.
[3] Zhang, Q. (2026). How to Improve Marketing Efficiency and Precision through AI-Driven Innovative Products. Strategic Management Insights, 3(1), 1-8.
[4] Liu, Y. (2026). The Promoting Role of Fintech and Product Innovation in the Context of the Digital Economy. Strategic Management Insights, 3(1), 9-16.
[5] Cai, Y. (2026). Design and Implementation of System Extensibility under High Concurrency Environment. International Journal of Engineering Advances, 3(1), 31-37.
[6] Liu, Y. (2026). The Application of Data-Driven Financial Risk Management in Multinational Enterprises. Economics and Management Innovation, 3(1), 20-26.
[7] Huang, J. (2026). Practice of Public Space Optimization and Functional Enhancement in Cultural Architecture. European Journal of Engineering and Technologies, 2(1), 9-21.
[8] Xu, D. (2026). AI-Driven Video Content Optimization Strategies for Immersive Media. European Journal of Engineering and Technologies, 2(1), 1-8.
[9] Qi, Y. (2026). AI Driven Payment System Security Improvement and User Privacy Protection Mechanism. Journal of Computer, Signal, and System Research, 3(1), 35-41.
[10] Sun, J. (2025). Research on Financial Systemic Risk Measurement Based on Investor Sentiment and Network Text Mining. Socio-Economic Statistics Research (2025), 6(2), 185-193.
[11] Lu, Z. (2025). Design and Practice of AI Intelligent Mentor System for DevOps Education. European Journal of Education Science, 1(3), 25-31.
[12] Zhang, X. (2025). Optimization and Implementation of Time Series Dimensionality Reduction Anti-fraud Model Integrating PCA and LSTM under the Federated Learning Framework. Procedia Computer Science, 262, 992-1001.
[13] Zhang, X. (2025, May). Automobile Finance Credit Fraud Risk Early Warning System based on Louvain Algorithm and XGBoost Model. In 2025 3rd International Conference on Data Science and Information System (ICDSIS) (pp. 1-7). IEEE.
[14] Jin Li. Performance Analysis of Efficient Microservice Architecture in the Financial Industry. Machine Learning Theory and Practice (2026), Vol. 6, Issue 1: 1-9.
[15] Yixian Jiang. Performance Optimization and Improvement of Advertising Machine Learning Platform Based on Distributed Systems. International Journal of Big Data Intelligent Technology (2026), Vol. 7, Issue 1: 9-17