基于注意力机制的交互式神经网络模型在细粒度情感分类中的应用  

APPLICATION OF INTERACTIVE NEURAL NETWORK MODEL BASED ON ATTENTION MECHANISM IN FINE-GRAINED SENTIMENT CLASSIFICATION

作　　者：杨鹏 杨青 李志斌[2] 王扬 Yang Peng;Yang Qing;Li Zhibin;Wang Yang(Information Communication Company,State Grid Tianjin Electric Power Company,Tianjin 300010,China;State Grid Tianjin Electric Power Company,Tianjin 300010,China)

机构地区：[1]国网天津市电力公司信息通信公司,天津300010 [2]国网天津市电力公司,天津300010

出　　处：《计算机应用与软件》2020年第7期130-135,共6页Computer Applications and Software

基　　金：天津市科技计划项目(18ZXZNGX00310);国网天津市电力公司科技项目(kj18-1-17)。

摘　　要：细粒度情感分类是情感分析领域中一个重要的任务。给定意见句和方面词,该任务的目标是识别该意见句关于给定方面词的情感极性。情感极性不仅与上下文语义相关,而且与方面词的语义也有密不可分的联系。现有的研究工作通常仅利用方面词语义来生成特定的上下文表示,忽略了对方面词本身的语义建模。针对该问题,提出一个基于注意力机制的交互式神经网络模型,可以同步进行上下文语义和方面词语义的交互式建模,进而更好地生成上下文和方面词的表示,提升情感分类的效果。分别在两个公开数据集上进行实验,并与现有的方法进行比较。实验结果表明,该模型取得了最好的性能。Fine-grained sentiment classification is an important task of sentiment analysis field.Given an opinion sentence and an aspect term,the goal is to identify the sentiment polarity of the given aspect term.Therefore,both of the semantics of context and aspect term are related to the sentiment polarity.The existing works usually only use aspect term to generate specific context representation,and ignore the semantic modeling of aspect terms themselves.To solve this problem,this paper proposes an interactive neural network model based on attention mechanism.It could be used for interactive modeling of context semantics and aspect terms simultaneously,so as to better generate the representation of context and aspect terms and improve the effect of sentiment classification.We experimented on two open datasets and compared them with the existing methods.The experimental results show that the model achieves the best performance.

关 键 词：细粒度 情感分类 交互式 注意力机制 

分 类 号：TP3[自动化与计算机技术—计算机科学与技术]