摘要
在保险领域智能问答应用研究中,用户提问时大量使用缩写、简写的保险名称,降低了问题语义理解的准确率。为解决这个问题,本文提出使用条件随机场(CRF)与双向长短记忆循环神经网络相结合的模型(Bi-LSTM-CRF),加入预先训练好的字嵌入向量进行训练的方法来识别保险名称。实验结果表明,CRF结合双向的LSTM的方法相较于传统机器学习的方法,在保险领域命名实体的识别中具有更好的性能,显著提高了保险名称识别的准确率和召回率。
Because the insurance name is long and there are a lot of abbreviations and ambiguities in the user's query,identifying the insurance naming information in the user question becomes a research problem in insurance intelligent question answering.In this paper,a model(Bi-LSTM-CRF) combining Conditional Random Field(CRF) with two-way long-term memory networks(BiLSTM) is proposed.Then pre-trained word embedding vectors are added to train this model.The experiment result shows that:compared with the traditional method of machine learning,Bi-LSTM-CRF has a better performance in insurance Name Entity Recognition.
作者
陈彦妤
杜明
CHEN Yanyu;DU Ming(School of Computer Science and Technology,Donghua University,Shanghai 201620,China)
出处
《智能计算机与应用》
2018年第3期111-114,共4页
Intelligent Computer and Applications