Chinese named entity recognition in the field of Architecture

摘要

This paper presents an IDCNN-BILSTM-CRF model for named entity recognition in Chinese architecture field. This model uses the text data in the field of architecture constructed in this paper. First, the word vectors are put into the IDCNN model for feature extraction, then the accuracy of feature extraction is further improved in the input BILSTM model, and finally the CRF is used to find the optimal solution of annotation. And in order to greatly reduce the training speed, this paper also puts forward feasible solutions.