韩忠明(1972-),男,山西文水人,博士,副教授,CCF专业会员,主要研究领域为社会网络,数据挖掘,大数据处理;张梦玫(1995-),女,硕士生,主要研究领域为自然语言处理;李梦琪(1993-),女,硕士生,主要研究领域为自然语言处理;段大高(1976-),男,博士,副教授,CCF专业会员,主要研究领域为数据挖掘,人工智能;刘雯(1992-),男,硕士生,主要研究领域为社交网络挖掘;于重重(1971-),女,博士,教授,CCF高级会员,主要研究领域为模式识别,机器学习.
韩忠明,E-mail:hanzhongming@btbu.edu.cn
国家自然科学基金(61170112,61532006);北京市自然科学基金(4172016)
National Natural Science Foundation of China (61170112, 61532006); Beijing Natural Science Foundation (4172016)
网络评论的观点挖掘任务是文本分析的关键问题之一.随着网络评论的快速增长,用户在浏览评论时更加关注细粒度的信息,因此,对评论进行方面级观点挖掘能够帮助消费者更好地做出决策.过去的10多年间,研究人员在大量网络评论语料库上进行观点挖掘等相关研究,并取得了丰硕的研究成果和广泛的应用价值,更不乏优秀学者对观点挖掘方法现状进行综述总结.然而,有针对性地对观点挖掘中的方面提取与观点提取进行综述总结的成果较少.综述了近年来网络评论方面级观点挖掘的研究现状:首先,介绍了方面级观点挖掘的相关问题描述;然后,重点分类介绍方面提取方法及观点内容提取的主要方法;随后,总结了方面级观点挖掘的常见评价指标以及在社会中的广泛应用价值;最后,根据对现有方法提出具有挑战性的方向并进行系统总结.对方面级观点挖掘进行综述有助于比较不同方法的差异,从而发现有价值的研究方向.
Opinion mining (OM) of Internet reviews is one of the key issues in text analysis. As the rapid growth of the Internet reviews, users pay more attention to all this fine-grained information when browsing comments. Therefore, aspect-level OM can help consumers make better decisions. In last decade, researchers conducted opinion extraction and analysis on a large number of Internet reviews corpus, and have achieved fruitful research results and broaden the scope of application. There were also some scholars conducted summaries on the present situation of OM methods. To rectify the lack of specific summaries on aspect extraction and opinion expression extraction, this paper analyzes and summarizes the recent research status of aspect-level OM on Internet reviews. The paper describes the aspect-level OM, introduces the different methods of aspect extraction and opinion expression extraction, and summarizes the evaluation measures of aspect-level OM and application values. In the end, it provides an overview of the future challenges along with a synopsis on the existing techniques. This specific survey on aspect-level OM helps to evaluate the different methods and find valuable research direction.
韩忠明,李梦琪,刘雯,张梦玫,段大高,于重重.网络评论方面级观点挖掘方法研究综述.软件学报,2018,29(2):417-441
复制