基于方面语义和门控过滤网络的方面级情感分析  

Aspect-based Sentiment Analysis Based on Aspect Semantic and Gated Filtering Network

作　　者：何智豪 陈红梅[2] 罗川 HE Zhihao;CHEN Hongmei;LUO Chuan(Tangshan Research Institute,Southwest Jiaotong University,Tangshan,Hebei 063000,China;School of Computing and Artificial Intelligence,Southwest Jiaotong University,Chengdu 611756,China;College of Computer Science,Sichuan University,Chengdu 610065,China)

机构地区：[1]西南交通大学唐山研究院,河北唐山063000 [2]西南交通大学计算机与人工智能学院,成都611756 [3]四川大学计算机学院,成都610065

出　　处：《计算机科学》2023年第10期193-202,共10页Computer Science

基　　金：国家自然科学基金(61976182,62076171);四川省自然科学基金(2022NSFSC0898)。

摘　　要：方面级情感分析(Aspect-based Sentiment Analysis,ABSA)是一项细粒度的情感分析任务,旨在预测文本中特定方面的情感极性。目前,鉴于循环神经网络在序列建模方面的卓越性能以及卷积神经网络学习局部模式的出色表现,部分工作将两者相结合来挖掘情感信息,并且取得了不错的效果。但是,少有工作在将两者结合后应用到方面级情感分析任务中的同时考虑方面信息。在方面级情感分析任务中,大部分工作将方面视作一个独立整体与上下文进行交互,但是对于方面的表示过于简单,缺乏真实语义。针对上述问题,文中提出了一种基于方面语义和门控过滤网络(Aspect Semantic and Gated Filtering Network,ASGFN)的神经网络模型,用于挖掘方面级情感信息。首先,设计了方面编码模块,用于捕捉特定语境下的方面语义信息,该模块基于全局上下文融合多头注意机制与图卷积神经网络构建包含特定语义的方面表示。其次,设计门控过滤网络连接循环神经网络和卷积神经网络,以此增强方面与上下文的交互,同时结合循环神经网络与卷积神经网络的优势,进而提取情感特征。最后,将情感特征与方面表示相结合,生成预测情感极性的语义表征。在restaurant,laptop和twitter这3个公用数据集上分别取得了84.72%,78.64%,76.22%的情感分类准确率。实验结果表明了所提模型的有效性,它能提高方面级情感分类任务的性能。Aspect-based sentiment analysis(ABSA)is a fine-grained sentiment analysis,which aims to predict sentiment polarity of text toward a specific aspect.Currently,given the excellent capabiities of recurrent neural networks(RNN)in sequence modeling and the outstanding performance of convolutional neural networks(CNN)in learning local patterns,some works have combined the two to mine sentiment information and achieved good results.However,few works consider aspect information while applying the combination of the two to ABSA.In aspect-based sentiment analysis tasks,most of the work treat aspect as an independent whole interacting with the contexts,but the representation of aspect is too simple and lacks real semantic.To address the above issues,this paper proposes a neural network model based on aspect semantic and gated filtering network(ASGFN)to better mine aspect-based sentiment information.First,an aspect encoding module is designed to capture context-specific aspect semantic information,which is based on a global context fusion multi-head attention mechanism with a graph convolutional neural network to construct aspect representation containing specific semantic.Second,a gated filtering network is designed to connect RNN and CNN as a way to enhance the interaction of aspect with the contexts,while combining the advantages of the RNN and the CNN,and then extracting the sentiment feature.Eventually,the sentiment feature is combined with aspect representation to generate semantic representation that predicts sentiment polarity.Sentiment classification accuracies of 84.72%,78.64%,and 76.22%are achieved in three communal datasets,restaurant,laptop,and twitter,respectively.Experimental results demonstrate the effectiveness of the proposed model,which can improve the performance of ABSA.

关 键 词：方面级情感分析 方面语义 门控过滤 循环神经网络 卷积神经网络 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]