摘要
在线评论所包含的产品特征观点在帮助用户做出购买决策时发挥着重要作用,然而,目前还没有挖掘客户评论中的产品特征观点作为主要数据来源的推荐系统,并且,监督型方法中提取特征观点对的算法较少关注中文句式结构,提取规则缺乏动态适应性.因此,提出一种基于特征观点对的产品推荐模型,首先,结合中文句式构成分析及特征观点的匹配关系分析,采用动态窗口提取特征观点对;在此基础上,结合特征树汇聚特征观点用于产品间比较,并为用户做出产品推荐;同时,提出情感可信度指标用于展示特征的典型评论.与采用静态窗口的基准方法相比,本模型的召回率和F值都有大幅提升,表明其可以为基于特征观点对的产品推荐提供可靠的数据来源,进而有效帮助用户做出购买决策.
The product feature opinions included in online customer reviews play an important role in helping users make purchase decisions. However, there is no recommendation system which mines product feature opinion from customer comments as the main data source. Furthermore, the feature opinions extraction algorithms of supervised methods focus less on Chinese sentence structure, and the extraction rules are lack of dynamic adaptability. Therefore, this paper proposes a product recommendation model based on feature opinion pairs. Firstly, combining the analysis of Chinese sentence structure and the matching relationship of feature opinion, the dynamic window is used to extract feature opinion. Secondly,aggregating feature opinions based on feature tree for product comparison and product recommendation.Finally, proposing the indicator of emotional credibility for typical review display. Compared with the reference methods using static window, our model has high recall and F-score in feature opinion extraction,which shows that it can provide reliable data sources for product recommendation based on feature opinion pairs, thus help users make purchase decisions effectively.
作者
郝玫
马建峰
HAO Mei;MA Jianfeng(Donlinks School of Economics and Management,University of Science and Technology Beijing,Beijing 100083,China)
出处
《系统工程理论与实践》
EI
CSSCI
CSCD
北大核心
2018年第9期2363-2375,共13页
Systems Engineering-Theory & Practice
基金
北京市社会科学基金(17GLC061,16LJB002)~~
关键词
在线客户评论
文本挖掘
特征观点对
产品推荐
动态窗口
online customer reviews
text mining
feature opinion pair
product recommendation
dynamic window
