Bi-LSTM-based legal named entity identification method

The invention discloses a legal named entity recognition method based on Bi-LSTM, and belongs to the field of natural language processing. According to the method, word vector training and a deep learning method are combined, and a named entity recognition task in the field of natural language processing is deeply improved and optimized, so that the problems of accuracy and model complexity are considered. The method comprises the steps of firstly performing text preprocessing to reduce a large amount of interference information in original text data; then word vector training is carried out on the processed corpus, and word vectors are trained based on a skip-gram model; the invention provides a method for completing corpus feature extraction by utilizing Bi-LSTM, and the problem of named entity recognition in the legal field is solved by combining the relationship between CRF limitation tags and performing result correction. 本发明公开了一种基于Bi-LSTM的法律命名实体识别方法，属于自然语言处理领域。本发明结合词向量训练和深度学习方法，对自然语言处理领域的命名实体识别任务进行了深度的改