基于加权频繁子树相似度的网页评论信息抽取-计算机应用研究.PDFVIP

优先出版 计 算 机 应 用 研 究 第33 卷 基于加权频繁子树相似度的网页评论信息抽取* 1, 2 1 1 1 1 郝志峰 ，袁 琴 ，蔡瑞初 ，温 雯 ，骆魁永 (1．广东工业大学 计算机学院，广州 510006 ；2 ．佛山科学技术学院，广东 佛山 528000) 摘要快速积累的海量产品评论信息是商家和消费者进行需求调研或购物决策时的重要依据。针对现有网页信息抽取方法 普遍存在人工耗时大，抽取准确率低等问题，提出了一种基于加权频繁子树相似度的网页评论信息抽取方法 WTS 。首 先通过视觉特征对网页进行剪枝处理。然后，通过深度加权的相似度度量方法抽取最佳频繁子树。最后，通过子树对齐 方法抽取评论路径并解析评论内容。通过对京东、苏宁等网站的评论内容抽取实验，验证了WTS 比D-EEM、POL 等 方法在抽取产品评论信息上具有一定的优势。 关键词 产品评论；抽取；加权频繁子树 中图分类号：TP391.1 Page comments extraction based on weighted similarity of frequent sub-tree 1, 2 1 1 1 1 Hao Zhifeng , Yuan Qin , Cai Ruichu , Wen Wen , Luo Kuiyong (1. College of Computer Science, Guangdong University of Technology, Guangzhou 510006, China; 2. Foshan University, Foshan Guangdong 528000, China) Abstract: Massive product comments are accumulated, which is important information for vendor or consumer in the process of selling or buying. Aiming at the problem that existing method is time consuming and poor performance. This paper proposed WTS method based on weighted frequent sub-tree similarity, which could extract product comments information rapidly and accurately. Firstly, Web pages were pruned through visual characteristics. Then, the best frequent sub-tree was extracted by depth-weighted similarity measure. Lastly, comment path was extracted and comment content was parsed . Extracting product comments from the website of JD, SN and so on, experimental results show that WTS has better performance