计算机工程与应用2025，Vol.61Issue(19)：60-73,14.DOI:10.3778/j.issn.1002-8331.2409-0258
方面级情感分析研究综述
Survey on Aspect-Based Sentiment Analysis
摘要
Abstract
With the continuous development of the sentiment analysis,aspect-based sentiment analysis(ABSA)is pro-posed,which is an analytical method that focuses on analyzing the evaluated object and its attributes.ABSA has received increasing attention.Compared with traditional sentiment analysis,ABSA can accurately reflect the evaluated aspect and its opinions,helping enterprises make targeted improvements and promote high-quality development.According to different methods,they can be divided into rule-based methods,attention based methods,machine reading comprehension based methods,sequence annotation based methods,and generative based methods.Existing research is analyzed and summa-rized.In addition,this paper also introduces new tasks in ABSA:multimodal,implicit,and cross-domain aspect-based sen-timent analysis;provides common data resources such as SemEval series,MAMS,ACOS,and CMU-MOSEI.Finally,this paper explores the challenges faced and future research directions,such as small data volumes,single domains,and une-valuable data quality.It emphasizes the importance of enhancing aspect-based sentiment analysis for complex sentences and multimodal contexts,and explores how to better leverage pre-trained language models to provide insights for more in-depth research.关键词
机器学习/深度学习/方面级情感分析Key words
machine learning/deep learning/aspect-based sentiment analysis分类
信息技术与安全科学引用本文复制引用
徐放,尹伊纯,吴方君..方面级情感分析研究综述[J].计算机工程与应用,2025,61(19):60-73,14.基金项目
江西省教育厅科技项目(GJJ210508). （GJJ210508）
