Entity Recognition for Military Situation Awareness Knowledge Graph with Wikipedia Data

Entity recognition is an essential component of knowledge representation and knowledge extraction research. To enhance military situation awareness through the construction of a knowledge graph, this paper presents a novel method, BERTATT_POSBiLSTMLSTMCRF, which is based on the traditional entity recognition model BERT_BiLSTM_CRF. The local location information and the impact of the entity's position in the sentence on the entity recognition task are both fully considered by introducing the attention mechanism. Additionally, an LSTM layer is added after the BiLSTM layer to deal with long-distance label dependencies while improving the model's ability to recognize long entities. Comparative experiments demonstrate that the improved model proposed in this paper is effective in entity recognition with Wikipedia data.