Military Named Entity Recognition Method Based on Deep Learning

Abstract

Military information extraction is an important means to gain military information advantage. Military named entity recognition is the basis of military information extraction. For military named entities, a method of supervised named entity recognition based on deep learning was proposed to identify and extract military named entities in military texts such as troops, geographical locations, weapons and so on. This method avoids the complexity of the artificial construction of features and inaccuracy of military text segmentation. It uses BI-LSTM and CRF model to automatic extract text features based on character embedding, and then identifies military named entities. Experiments show that this method is inferior to traditional methods, and is close to the level of named entity recognition in the general field.