基于目标对象表示学习的立场判定  

Standpoint Determination Based on Target Object Representation Learning

作　　者：童晓薇[1] 曾思通[1] 林潇丽[1] TONG Xiaowei;ZENG Sitong;LIN Xiaoli(Department of Mechanical Engineering,Fujian Chuanzheng Communications College,Fuzhou 350007,China)

机构地区：[1]福建船政交通职业学院机械工程系,福建福州350007

出　　处：《新乡学院学报》2020年第9期20-24,共5页Journal of Xinxiang University

摘　　要：针对基于文本的立场判定通常会被归结为简单分类或情感分析而出现判定不准确的问题,提出一种基于目标对象表示学习的文本立场判定模型。模型能充分利用卷积神经网络在局部特征提取方面的优势,实现在不依赖复杂的句法树解析过程的同时,识别决定特定立场分析的关键目标,有效提升文本立场判定模型性能。以微博中针对某大型央企产品及政策的舆论文本为例,开展立场判定实验,通过与多个主流基线模型的比较,验证了方法的有效性和可行性。Aiming at the problem that text-based position determination is usually attributed to simple classification or sentiment analysis which often leads to inaccurate determination,a text position determination model based on target object representation learning was proposed.The model can make full use of the advantages of the convolutional neural network in local feature extraction,and achieve the identification of key goals that determine the specific position analysis without relying on the complex parsing process of the syntactic tree.It can effectively improve the performance of the text standpoint determination model.Taking the texts of public opinion on the products and policies of a large state-owned enterprise in Weibo as an example,the standpoint determination experiment is carried out.The effectiveness and feasibility of the method are verified by comparing with multiple mainstream baseline models.

关 键 词：立场判定 目标对象表示学习 深度学习 自然语言处理 卷积神经网络 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]