民航突发事件实体识别方法研究

针对民航突发事件实体自动获取的问题,提出一种将双向长短期记忆网络(BiLSTM)与条件随机场(CRF)结合的民航突发事件实体识别方法。利用领域词典对民航突发事件实体进行自动标注,并将已标注的数据转化为分布式字符向量;采用BiLSTM模型获取序列文本的上下文特征;通过CRF模型获得序列标注结果。实验结果表明,与传统的CRF方法相比,该方法的F值有明显的提升,有效解决了传统方法在识别民航突发事件文本信息中包含复合实体或混合实体方面效率低且大量使用人工定义特征模板的问题,为民航突发事件知识图谱的自动构建提供了更好的方法支撑。

ENTITY RECOGNITION METHOD OF CIVIL AVIATION EMERGENCY

Aiming at the problem of automatic acquisition of civil aviation emergency entities,this paper proposes a method of civil aviation emergency entity recognition based on the combination of BiLSTM and CRF.We used the domain dictionary to automatically label the civil aviation emergency entities,and transformed the labeled data into distributed character vectors.BiLSTM model was adopted to get the context features of sequence text,and CRF model was used to get the sequence labeled results.The experimental results show that compared with the traditional CRF method,the F-value of this method is significantly improved,which effectively solves the problem that the traditional method is inefficient in identifying composite or mixed entities in the text information of civil aviation emergencies and uses a large number of manually defined feature templates.And it provides a better method support for the automatic construction of civil aviation emergencies knowledge graph.