A Method for Resume Information Extraction Using BERT-BiLSTM-CRF
Xiaowei Li, Hui Shu, Yi Zhai, Zhiqiang Lin
Abstract
To solve the problem of low efficiency of electronic resume information extraction by artificial construction rules, a resume information extraction method based on named entity recognition is proposed, which extracted personal details such as graduation college, job intention and job skills from the resume into named entity recognition. Firstly, the TXT text in different formats of resume file is extracted for data cleaning and other preprocessing. The BERT language model based on multi-head self-attention mechanism is used to extract text features and obtain word granularity vector matrix. The BiLSTM neural network is used to obtain the context abstraction features of serialized text. Finally, using CRF to decode and annotate the global optimal sequence, the corresponding resume entity information is extracted. Experimental results show that the whole scheme can effectively extract electronic resume information, and the performance of the resume information extraction model based on BERT-BiLSTM-CRF is better than other models.