用户评论标签摘要系统的有效性研究

The Effectiveness of Tag-based Review Summarization Systems

信息过载是消费者处理用户评论信息时面临的一个巨大挑战。由于评论数量超过了个人信息处理能力,消费者不得不借助某些启发规则(如重点关注中评、差评以及最近发表的评论)来简化阅读评论的任务,在降低了认知负载的同时,也增加了用户产生选择性偏见的可能。基于标签的评论摘要是电子商务网站新出现的一种工具,将评论按照产品属性或用户感受分类,并且给每类一个标签。这一工具对缓解用户信息过载和选择性偏见有很大的潜力,但是其作用机理和使用效果还未得到实证验证。以图式理论和认知适应理论为基础,针对两种不同类型的商品(体验品和搜索品),研究了基于标签的摘要对系统感知有用性和满意度的影响。实验结果表明:对于体验类商品,与传统在线评论展现方式相比,增加基于标签的摘要不能提高感知有用性和满意度;然而,对于搜索类商品,与传统在线评论展现方式相比,增加基于标签的摘要能提高用户感知有用性和满意度。

Online review is becoming a major factor driving consumer purchase decisions.However,review information overload is a big challenge for consumers.As the number of reviews is overwhelmingly beyond a person＇s information processing capability,the consumers have to turn to some heuristics（e.g.,focus only on neutral,negative or latest comments）to simplify the task of comments reading.This reduces the cognitive load but also increases the possibility of selectivity bias.Tag based review summarization is a tool emerging on e-commerce website,which classifies customer reviews into categories by product attributes or user perceptions,and assigns a label for each category.Although the tool is promising to alleviate information overload and selectivity bias,its mechanism and effectiveness has not been studied.Based on schema theory and cognitive fit theory,we investigate the impact of tag based review summarization on two types of products（i.e.,experience product and search product）in terms of perceived usefulness and system satisfaction.The results indicate that for experience products,compared with traditional review presentation,providing tag based summarization does not lead toimproved perceived usefulness and system satisfaction.However,for search products,compared with traditional review presentation,providing tag based summarization leads to improved perceived usefulness and system satisfaction.