Semi-markov CRF Model Based on Stacked Neural Bi-LSTM for Sequence Labeling
Yupeng Liu, Guodong Li, Xiaochen Zhang · 2020 IEEE 3rd International Conference of Safe Production and Informatization (IICSPI) · 2020
In this paper, we present a model for sequence labeling. Based on word vector representations of neural network model algorithms and artificial features of statistical machine learning, a stack Bi-LSTM and new semi-Markov conditional random field frameworks are proposed. On the one hand, we present stack Bi-LSTM to achieve segmentation and chunking task on both word-level and character-level. On the other hand, we use traditional CRF and NSCRF to make some restrictions on the labeling entities to improve the accuracy of labels and then use them to decode the unknown sequence. We test the trained model in the CoNLL2000 and CoNLL2003 datasets.