NER Based on feed-forward depth neural network
Aiping Xu, Chao Wang · 2021 International Conference on Computer Information Science and Artificial Intelligence (CISAI) · 2021
The named entity recognition technology refers to the identification of human names, place names, institutions and other nouns from the text, is the key basic work of the information extraction, question-and-answer system, machine translation and other applications, its research is in full swing. In this paper, based on feed-forward depth neural networks, in the method of window, the named entities in text are identified and a series of comparative experiments are carried out. The experiments show that in the named entity recognition the reasonable settings for the window, the replacement processing of low-frequency words and the use of BIO tags can improve the entity recognition performance of the feed-forward depth neural network model. The data comparison basis is obtained for subsequent research on the identification of named entities based on LSTM depth neural network.