In the financial field,the complexity of professional terminology and the dependence between sentences lead to a lower accuracy of causality extraction.To address this issue,a causality extraction model based on BERT semantic enhancement was proposed,including a basic model and an enhanced model,to obtain rich text features and achieve semantic deep extraction.The BERT pre-training model was used to obtain contextual features,and the adversarial learning of the adversarial neural network was employed to further learn high-discriminative features,thus improving the accuracy of causality extraction.Experimental results demonstrate that the proposed model can improve the accuracy of causality extraction.
关键词
因果关系抽取/信息抽取/金融领域/对抗神经网络/对抗学习/基本模型/增强模型
Key words
causality extraction/information extraction/financial field/anti-neural network/confrontational learning/basic model/enhanced model
维普
万方数据