基于深度学习的中文网络招聘文本中的技能词抽取方法

Skill word extraction from Chinese online recruitment texts based on deep learning

为了能够充分利用领域知识来提升技能词的抽取性能,提出了一种基于深度学习与语料特征相结合的技能词抽取方法。将技能词抽取转化为序列标注问题,以序列标注的基本模型Bi-LSTM-CRF为基础,在输入层中加入语料特征,并将输入层的输出与Bi-LSTM输出连接在一起作为CRF层的输入。实验结果表明,提出的技能词抽取方法效果提升明显,加入的语料特征有利于提升技能词抽取的准确率,并能够缓解标注数据的稀缺。

In order to make full use of domain knowledge to improve the performance of skill word extraction,a method of skill word extraction based on the combination of deep learning and corpus features is proposed.Skill word extraction is transformed into a sequence tagging problem.Based on the basic model of sequence tagging,Bi-LSTM-CRF,which adds corpus features to the input layer,and connects the output of the input layer with the output of the Bi-LSTM layer as the input of the CRF layer.The results of a large number of experiments show that the effect of the proposed skill word extraction method has improved significantly.The added corpus features can help improve the accuracy of skill word extraction and reduce the effort of data annotations.