融合上下文信息的隐式情感句判别方法  

An Implicit Sentiment Sentence Identification Method Base on Context Information

作　　者：王素格[1,2] 王敏[1] 廖健[1] 陈鑫[1] WANG Suge;WANG Min;LIAO Jian;CHEN Xin(School of Computer and Information Techmology,Shanxi Univrsit,Taiyuan 030006,China;Key Laboratory of Computational Intelligence and Chinese Information Processing of Ministry of Education,Shanxi University,Taiyuan 030006,China)

机构地区：[1]山西大学计算机与信息技术学院,山西太原030006 [2]山西大学计算智能与中文信息处理教育部重点实验室,山西太原030006

出　　处：《山西大学学报（自然科学版）》2022年第2期380-386,共7页Journal of Shanxi University(Natural Science Edition)

基　　金：国家自然科学基金(61906112;62076158);山西省应用基础研究计划项目(201901D211174);山西省高等学校科技创新项目(2019L0008;2020L0001);山西省重点研发计划项目(201803D421024)。

摘　　要：文本情感分析作为自然语言处理领域中热点研究方向之一受到持续的广泛关注。相较于显式情感表达,隐式情感表达本身所提供的信息不足。文章提出一种融合上下文信息的隐式情感句判别方法。使用多极性正交注意力机制对隐式情感句嵌入表示,建立融合上下文的多极性注意层,对上下文信息进行建模,以挖掘上下文中的关键信息。最后,将上下文信息表示与隐式情感句表示拼接,以弥补隐式情感句本身信息不足的问题。在SMP2019-ECISA隐式情感分析数据集上的实验结果表明,该模型充分利用了上下文信息,提升了隐式情感句判别的效果。As one of the hot research directions in the field of natural language processing,text sentiment analysis has received continuous and extensive attention.Compared with explicit sentiment expression,the information provided by the implicit sentiment expression itself is insufficient.This paper introduced a method for identify implicit sentiment sentence with contextual information.Expressions of implicit sentiment sentence are embedded by using a multi-polarity orthogonal attention representation model,and a multi-polarity attention layers are established by integrating context and the model the context information to mine the key information in the context.Moreover,the contextual information representation and the implicit sentiment sentence representation are spliced together to make up for the lack of information in the implicit sentiment sentence itself.The experimental results on the SMP2019-ECISA implicit sentiment analysis dataset show that the proposed model makes full use of context information and improves the effect of implicit sentiment sentence identification.

关 键 词：隐式情感分析 上下文信息 多极性正交注意力 

分 类 号：TP391[自动化与计算机技术—计算机应用技术]