消费咨询网络中的虚假观点识别与产品采用研究

消费咨询网络是社会化商务环境中的一类特殊社交网络，商家借助消费咨询网络用户传播产品信息从而形成了社交广告。当网络中传播的社交广告传递高于产品品质的评价，或对竞争对手的产品做出低于真实产品品质的评价，称为虚假评价。基于连续观点和离散行为思想构建了产品采用模型，并在模型中设置了用户观点学习规则，即尚未购买产品的用户优先选择诚信指数较高的用户进行观点学习。借助计算实验方法研究了社交广告、观点学习轮次对观点演化和产品采用的影响。结果表明：产品采用模型能够使得被虚假观点误导从而购买产品的用户比例降低；虚假观点识别机制可以有效识别出"低质高评"和"高质低评"，并降低其导向性；观点学习轮次的增加可以提高虚假观点的识别效果。研究克服了现有产品扩散模型在不区分信息扩散与产品购买、不考虑信息真伪性、不考虑不同采用阶段等方面的局限。

Untruthful Opinions Idenfication and Product Adoption in Consumer Advice Network

The consumer advice network is a special type of social network in a social commerce environment. Users are able to establish online social relationship and sharing product evaluation, and search product related information from network members. As merchants start to spread product information through influential users in the consumer advice network, this activity lead to the emergence of social advertising, or user-generated advertising. In this background, merchants may rely on external incentives to stimulate uses to post and diffuse social advertising on the consumer advice network. By using social advertising tools, merchants make higher-than-true product quality evaluations for their products, or make lower-than-true product quality evaluations for competitors' products. Such deceptive evaluations are called untruthful evaluations. Based on the concept of continuous opinion and discrete behavior, the model of product adoption is constructed, and the user opinion learning rules are set up in the model. That is, the users who have not yet purchased the product learning the opinion from those who have higher honest index. The influence of social advertising, opinion learning rounds on opinion evolution and product adoption are studied by means of agent-based computational experiment. The results show that the model can reduce the product adoption proportion of users who are misleaded by the untruthful opinions. The untruthful opinions identification mechanism can effectively identify the "low quality with high evaluation" and "high quality with low evaluation" and reduce the misleading effect from untruthful opinions. The increase of opinion learning rounds can increase the recognition of untruthful opinion. This Research overcome the limitations in existing product diffusion model, such as no distinction between information diffusion and product adoption, regardless of the authenticity of product information, regardless of the different stages of product adoption.