摘要
针对利用当前神经网络模型进行中文位置语义解析存在多义词解析效果差、泛化能力差等问题,提出一种基于BERT-BiLSTM-CRF模型的中文位置语义解析方法。首先利用BERT预训练模型对中文位置信息进行预训练,获取所有层中的上下文信息,增强中文位置信息的语义表征能力,然后通过BiLSTM模型提取向量特征信息,最后通过CRF模型进行解码,获取全局最优标注序列。实验结果表明,在不同数量和区域的中文位置信息数据集基础上,BERT-BiLSTM-CRF模型在所有测试集上分词准确率与F1值都优于目前常用的神经网络模型,最高分别可达到93.91%和93.96%。利用BERT-BiLSTM-CRF模型对中文位置信息进行语义解析,不仅有效提高了中文位置信息解析与多义词解析的准确率,而且具有更好的泛化能力。
Aiming at the problems of poor polysemous word parsing effect and poor generalization ability in Chinese location semantic analysis using current neural network models,propose a Chinese location semantic analysis method based on the BERT-BiLSTM-CRF model. This method first uses the BERT pre-training model to pre-train the Chinese location information to obtain the context information in all layers,enhance the semantic representation ability of the Chinese location information,and then extract the feature information of the vector through the BiLSTM model,and finally decode it through the CRF model to obtain the global Optimal labeling sequence. Experimental results show that on the basis of different numbers and regions of Chinese location information data sets,the word segmentation accuracy and F1 value of the BERT-BiLSTM-CRF model on all test sets are better than the current commonly used neural network models,and the highest F1 achieved are 93.91% and 93.96%,respectively. Using the BERT-BiLSTM-CRF model to perform semantic analysis on Chinese location information not only effectively improves the accuracy of Chinese location segmentation and polysemous word analysis,but also has better generalization capabilities.
作者
邓庆康
李晓林
DENG Qing-kang;LI Xiao-lin(Department of Computer Science and Engineering,Wuhan Institute of Technology;Hubei Key Laboratory of Intelligent Robot,Wuhan 430205,China)
出处
《软件导刊》
2022年第2期37-42,共6页
Software Guide
基金
国家重点研发计划项目(2017YFB0503701)
湖北省技术创新专项(2019AAA045)。
