新一代电子商务中面向用户生成内容的统计问题的研究

Web 2.0 technologies has seen rapid growth, the social interactions between customers and the firms has been very popular. Among all the various forms of social interactions, user-generated content (UGC) is undoubtedly the most prevalent one, which has fundamentally changed the way that information is created and diffused, as well as the consumer behavior and firm performance. Due to the large amount of data and other sorts of information it produces, UGC also brings more and more statistical questions into the frontier of the study field of e-commerce. This study seeks to address four statistis-related issues emerged in e-commerce field: 1) to compare existing sampling methods for the complex UGC data that is voluminous and inter-connected; 2)to construct a sampling method that can incorporate the indiviual attributes; 3) to investigate how to evaluate the unobsserved heterogeneity in the UGC data, especially in the online reviews; 4)to propose a new idea to design the online survey questionnaire, so that one can efficiently collect and analyze information from users directly, and update the questionnaire timely. These theories and methods we propose will help firms to improve their management and to enhance their competitive advantage by quickly,fully and correctly understanding their customers.

随着Web2.0技术的迅速发展，社会网络环境下用户与企业的互动变得十分普遍。用户创造内容（UGC）作为这种网络互动最主要的形式，其大量出现成为新一代电子商务的最重要特征，并对消费者行为和企业管理模式产生重要影响。有UGC所产生的海量数据也使得电子商务领域的研究面临很多与统计学相结合的问题。本课题拟从四个方面开展针对在线UGC衍生的统计学问题的研究: 1)在不同模式的社会网络环境下，具有海量交互的UGC数据的抽样方法比较；2）探索能够兼顾个体差异性的网络抽样方法；3）针对网络评价中的未被观测的异质性问题，提出纠偏的解决思路；4）对在线调查问卷的设计提出新的思路，以提高信息的采集，分析，问题更新的速度和有效性。通过对这些问题的探索，为企业更快速准确地了解用户提供理论支持和解决方案。

随着Web2.0技术的迅速发展，社会网络环境下用户与企业的互动变得十分普遍。用户创造内容（UGC）作为这种网络互动最主要的形式，其大量出现成为新一代电子商务的最重要特征，并对消费者行为和企业管理模式产生重要影响。由UGC所产生的交互型大数据也使得电子商务领域的研究面临很多与统计学相结合的问题。本课题从两个方面开展针对在线UGC衍生的统计学问题的研究: 1）对在线调查问卷的设计提出了新的思路，提高了信息的采集，分析，问题更新的速度和有效性。2）针对新兴的互联网金融行业，分析了在大数据环境下用户生成内容如何影响投融资者的行为模式。本研究课题从理论和实践两个角度，分别对互联网环境下用户生成内容的效率和有效性进行了推导和验证。