面向电子产品评论的比较关系挖掘-计算机软件与理论专业毕业论文.docxVIP

I I 摘 要 摘 要 从产品评论中获取信息的重要性，已经在生产、营销者和潜在的消费者中达成共识。 因此，产品评论挖掘也随之成为一个热门的研究课题。在这些产品评论中，存在着一定 数量的比较句。作为一种常见的表达方式，比较句可以帮助评论者方便地阐述两个产品 质量与服务的高低和优劣。因此，对这类评论进行研究值得重视。 本文对电子产品评论中的比较关系进行研究，主要工作包括以下几个方面： 在识别比较关系评论时，预先构建了比较关系特征词典和极比类型特征词典，并对 比较特征词按照极性进行分类，为后续工作做出准备。 将比较关系评论的识别看作一种分类问题，即“比较类”和“非比较类”。考虑到 比较关系评论通常具有固定的模式结构，因此首先采用类序列规则算法进行评论的一次 识别。针对一次识别中被判定为含有比较关系的评论和无法确定类别的评论，本文又给 出了基于产品命名规则的二次识别算法。依据评论中包含产品个数的不同，分别采用不 同的方法进行二次判断。 在抽取评论中的比较关系时，将抽取的结果表示为五元组。通过计算产品型号与网 页相关信息的相似度，完成比较对象的识别。而后构建出适用于比较关系评论的特征属 性抽取模式和结果值计算公式，依次完成五元组的抽取。 实验结果表明，利用本文的方法识别和抽取评论中的比较关系，能够提高准确率和 召回率。 关键词 评论挖掘 特征词典 类序列规则 命名规则 五元组 抽取模式 II II Abstract Abstract Customers and producers have already reached a consensus on the importance of getting information from product reviews. Product reviews mining has also become a hot research topic. A number of comparative reviews are included in product reviews. As a common expression, comparative sentences can help critics describe the difference of quality and service between two products. Therefore, the study on such reviews is important. This paper studies the comparative relations among electronic products reviews, and the main work includes the following aspects: When identifying the comparative reviews, a comparative feature lexicon and a superlative feature lexicon is firstly constructed. Then the study classifies the words in the comparative feature lexicon according to their polarities, laying a foundation for the follow-up. This study considers the identifying as a classification question, the “comparative” and “non-comparative”. Giving the comparative reviews usually following some fixed patterns, this study applies the class sequential rules to the first identification. For the candidate comparative reviews and the reviews without any classes after the first identification, a new identifying algorithm based on the product naming rules is proposed. Judging by the different numbers of products included in reviews, different methods are adopted for the se