基于CNN-BiLSTM-CRF的企业舆情监控模型构建  

Construction of enterprises public opinion monitoring model based on CNN-BiLSTM-CRF

作　　者：张欣艺 郑军红[1] 何利力[1] Zhang Xinyi;Zheng Junhong;He Lili(School of Computer Science and Technology,Zhejiang Sci-Tech University,Hangzhou,Zhejiang 310018,China)

机构地区：[1]浙江理工大学计算机科学与技术学院,浙江杭州310018

出　　处：《计算机时代》2023年第11期136-140,共5页Computer Era

基　　金：浙江省重点研发“尖兵”攻关计划项目(2023C01119)。

摘　　要：在互联网+环境下,将舆情监控应用到企业决策中是一个趋势。本文提出一种与企业需求契合度高且能更充分地提取情感特征的模型。将社交平台和互联网营销平台评论信息进行预处理,并使用Word2vec技术获取文本的词向量表示,通过CNN提取特征、BiLSTM提取上下文语义特征,再加入条件随机场模型对标签序列进行优化。最后使用K-means聚类得到企业网络舆论焦点。经在真实数据集上对比实验,验证了模型的有效性。Under the Internet+environment,it is a trend to apply public opinion monitoring to enterprise decision-making.A model that fits well with the needs of enterprises and can extract sentiment features more fully is proposed.The comment information of social platforms and Internet marketing platforms is preprocessed,and word vector representation of the text is obtained using Word2vec technique.The features are extracted by CNN,contextual semantic features are extracted by BiLSTM,and conditional random field model is added to optimize the label sequence.Finally,K-means clustering is used to obtain the enterprise online opinion focus.Comparison experiments on real datasets verify the effectiveness of the model.

关 键 词：企业舆情监控 CNN BiLSTM 舆论观点抽取 K-MEANS 

分 类 号：TP391.1[自动化与计算机技术—计算机应用技术]