Tibetan named entity recognition based on BiLSTM-CRF
Yangji JIA, Kunyu Qi, Cairang YUN · 2022
According to the characteristics of Tibetan named entity itself, this paper analyzes the text characteristics and recognition difficulties of named entity, and puts forward BiLSTM-CRF model. The combination of the two models not only makes use of the advantages of bidirectional LSTM model to save context information, but also reduces the influence of CRF layer from sentence level to consider before and after annotation, so as to achieve the effect of learning from each other's strengths and complementing each other's weaknesses, which is more effective to solve the problem of Tibetan named entity recognition. The experimental results show that the proposed recognition method has good performance, and the f-value can reach 88.7%.