【摘要】 近年来随着共享经济的迅猛发展,在线短租成为人们出行的最佳选择之一。但由于房东和房客之间存在着信息不对称问题,关系到房客的住宿安全和住宿体验,如何通过房客的评论内容判断最能影响后续房客对房源的信任度也日益成为文本挖掘领域和在线短租领域的研究热点。该文基于信任理论和不确定性减少理论,构建LDA主题模型,通过对主题特征词的维度标识确定每个主题的两级信任感知维度。最后通过实验结果分析和案例检验,验证了该文构建的LDA主题模型和信任维度划分的有效性,且不同房源空间分布会影响房源评论的主题分布,不同位置房源影响消费者对房源的信任值和关注度。更多还原

【Abstract】 With the rapid development of sharing economy, online short-term rent has become one of the best choices for people when traveling. However, due to the information asymmetry problem between landlords and tenants which is related to the tenants’ accommodation safety and experience, how to quantify the landlords’ credibility through the previous tenants’ comments has increasingly become a hot topic in the field of text mining and online short-term rent. Based on the theory of Trust and the theory of Uncertainty Reduction, this study marks the dimension of the topic feature tokens to make sure the two levels of trust-aware dimensions for each topic after constructing the Latent Dirichlet Allocation(LDA) theme model. In the end, this study proves the availability of the model and trust-aware dimensions division through the analysis for the experimental results and case test--it shows that spatial distribution of different rental houses has an impact on the theme distribution of reviews, and consumers show different trust value and concern on different houses as well.更多还原