|
|
一种基于FinBERT-CRF命名实体识别模型的证券领域知识图谱构建框架
|
Abstract:
随着信息媒介的转变以及人们对金融领域逐步的关注，证券领域新闻资讯信息的传递频率达到了前所未有的水平，而在当今金融领域缺乏一种能够可视化展示证券领域企业实体之间情感影响关系的建模方法。针对该问题，本文首先提出了一套实时的定向爬虫框架来获取所需的证券领域新闻文本，其次针对新闻文本设计了一种基于FinBERT-CRF的命名实体识别模型，最后结合市场基本面提出了一种构建面向情感分类的证券领域知识图谱，为投资者以及投资机构提供了一定的参考价值。
The frequency of news and information transmission in the securities has reached an unprece-dented level with people’s gradual attention to the financial domain and the transformation of in-formation media. However, there is no modeling method that can visually display the emotional impact relationship between corporate entities in the securities field. In response to this problem, this article first proposed a real-time directional crawler framework to obtain the required securi-ties field news text, then designed a FinBERT-CRF-based NER (named entity recognition) model for news. Finally, this paper proposes a knowledge map of securities field oriented to the sentiment classification based on the fundamentals of the market, which may provide a certain reference val-ue for investors and investment institutions.
| [1] | Devlin, J., Chang, M.W., Lee, K., et al. (2018) Bert: Pre-Training of Deep Bidirectional Transformers for Language Understanding. arXiv:1810.04805 [cs.CL] |
| [2] | Malkiel, B.G. and Fama, E.F. (1970) Efficient Capital Markets: A Review of Theory and Empirical Work. The Journal of Finance, 25, 383-417. https://doi.org/10.1111/j.1540-6261.1970.tb00518.x |
| [3] | http://choice.eastmoney.com/ |
| [4] | https://tushare.pro/ |
| [5] | Araci, D. (2019) FinBERT: Financial Sentiment Analysis with Pre-Trained Language Models.
arXiv:1908.10063 [cs.CL] |
| [6] | Panchendrarajan, R. and Amaresan, A. (2018) Bidirectional LSTM-CRF for Named En-tity Recognition. The 32nd Pacific Asia Conference on Language, Information and Computation (PACLIC 32), 2019. |
| [7] | Vaswani, A., Shazeer, N., Parmar, N., et al. (2017) Attention Is All You Need. arXiv:1706.03762 [cs.CL] |
| [8] | Wu, F., Liu, J., Wu, C., et al. (2019) Neural Chinese Named Entity Recognition via CNN-LSTM-CRF and Joint Training with Word Segmentation. The World Wide Web Conference, San Francisco, May 2019, 3342-3348.
https://doi.org/10.1145/3308558.3313743 |
| [9] | https://www.datafountain.cn/competitions/361 |
| [10] | López, F.M.S. and De La Cruz, E.G.S. (2015) Literature Review about Neo4j Graph Database as a Feasible Alternative for Replacing RDBMS. Industrial Data, 18, 135-139. https://doi.org/10.15381/idata.v18i2.12106 |
