Recommender and Reputation Systems: Principles and Practices

推荐系统和声誉系统：原则和实践

• 批准号: 0308006
• 负责人: Paul Resnick
• 金额: --
• 依托单位: Regents of the University of Michigan - Ann Arbor
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2003
• 资助国家: 美国
• 起止时间: 2003-07-01 至 2008-06-30
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=0308006&HistoricalAwards=false
• 关键词: Recommender; Reputation; Systems; Principles; Practices

Recommender and reputation systems process ratings from past users to guide current users' behavior. Such systems provide a natural laboratory for the study of large-scale social phenomena, including the provision of public goods (the ratings themselves), and the role of reputation as a regulator of trust and trustworthiness.This project will focus on two widely used systems, eBay and Slashdot. Usage logs provided by the two sites will be assessed in depth. This analysis should reveal the social practices that have emerged surrounding the provision of ratings, and the ways ratings actually guide users' decisions about what to buy and from whom, and what messages to read. Game-theoretic analyses of incentives will combine with the empirical analysis to guide the design of potential system improvements.This research should improve the functioning and public understanding of web sites where millions of people already interact with each other. In addition, it will provide fundamental insights into the workings of impersonal social capital, i.e., productive social relations that emerge among people who remain strangers rather than forming networks of long-term relationships.

推荐和信誉系统处理来自过去用户的评级以指导当前用户的行为。这类系统为研究大规模社会现象提供了一个天然的实验室，包括公共产品的提供（评级本身），以及声誉作为信任和可信度的调节器的作用。将深入评估这两个网站提供的使用日志。这一分析应该揭示围绕评级提供的社会实践，以及评级实际上指导用户决定购买什么、从谁那里购买以及阅读什么信息的方式。激励措施的博弈论分析将与实证分析联合收割机相结合，以指导潜在系统改进的设计。这项研究应该改善数百万人已经相互互动的网站的功能和公众的理解。此外，它将提供对非个人社会资本运作的基本见解，即，富有成效的社会关系出现在仍然是陌生人的人之间，而不是形成长期关系网络。