网络评论方面级观点挖掘方法研究综述

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.过去的十多年间，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方面提取与观点提取进行综述总结的成果较少.本文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；接着重点分类介绍方面提取方法及观点内容提取的主要方法；然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挑战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.

Survey of Studies on Aspect-Based Opinion Mining of Internet

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 these fine-grained information when browsing comments. Therefore, aspect-level OM can help consumers make better decisions. In last decade, the 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. Due to 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. This paper describes the aspect-level OM firstly, and then respectively introduces the different methods of aspect extraction and opinion expression extraction, then summarizes the evaluation measures of aspect-level OM and application values. Finally, we point out the future challenges and make a summary according to the exist techniques. The specific survey on aspect-level OM helps to evaluate the different methods and find valuable research direction.