Abstract: Relying on mobile Internet technology, user-generated content (UGC) platform has developed rapidly for the past years. UGC platform does not generate media content, but provides users with ways to generate and collaborate content, as well as distribution, customization and development of content methods, bilateral users are its core resources. With the support of information technologies such as big data algorithms and artificial intelligence, the user participation of UGC platform has witnessed explosive growth. By December 2020, the number of short video users in China had reached 873 million, with a usage rate as high as 88.3%. In the participation of a large number of waist and tail users, homogenization, vulgar and low-quality content problems have also gradually emerged.
UGC platform has been actively fulfilling its “gatekeeper” responsibility, through the examination and control of platform information content to achieve the purpose of governance. Since 2018, UGC platforms have been expanding their already large audit teams, and actively promoting the technical audit means of machine algorithm and artificial intelligence. However, in the context of the explosive growth of content quantity and richer content forms, high labor costs and the value defects of machine technology make passive audit more and more difficult. Actively promoting the active stimulation of quality content has become a more effective path of content governance. How to give full play to the importance of high-quality original content of head users to the content orientation, value establishment and user guidance of the platform, while balancing the increased subsidy cost and traffic loss caused by the platform to achieve the inspirational effect, has become an important direction of thinking to solve the dilemma of quality content cycle of UGC platform.
This paper studies the quality content generation system composed of head users and waist-tail users at the content generation end of UGC platform, and constructs three differential game models based on the platform subsidy head user driven decision, head users and waist-tail users collaborative decision, and decentralized decision without driving effect. In order to promote the ecological optimization of the content of UGC platform and realize the Pareto improvement of the overall revenue of the platform, this study selects the head users of the content generation end of UGC platform as the leading party of the game. Under the incentive of the platform subsidy policy, the driving strategy is adopted to form a Stackelberg differential game with the waist-tail users.
In this paper, we solve the respective revenue maximization objective function in the decentralized decision-making mode, the overall revenue maximization objective function in the collaborative decision-making mode and the two-stage objective function in the driven decision-making mode, respectively, to obtain the optimal strategy under different modes and the revenue and quality content level under the strategy. On this basis, the three modes are compared and analyzed, including the decision-making behavior, the respective income and the overall income, and some theoretical conclusions are obtained. The above conclusions are verified by using numerical example simulation technology. The example data come from the professional scores of 65 experts, scholars and senior users in related fields of industry and academia, and is verified and processed by orthogonal experiment. Finally, the relevant conclusions are obtained: (1)Under the head-driven effect, the high-quality content generation system of UGC platform has been significantly improved by Pareto. (2)the driving decision mode has a virtuous circle significance of improving the driving force of the head users. (3)The subsidy coefficient of UGC platform has a positive correlation with the improvement effect of quality content generation system, and the selection interval is controllable. The research conclusion provides a certain theoretical basis and practical guidance for UGC platform to effectively carry out content governance and promote the healthy ecological development of user-generated content.

Key words: UGC platform, high-quality content, differential game, platform subsidy, head users

摘要： 研究针对UGC平台内容生成端的头部用户和腰尾部用户所组成的优质内容生成系统,构建了基于平台补贴的头部用户带动决策、头部与腰尾部用户协同决策以及无带动效应的分散决策三种微分博弈模型,随后进行了结果对比与算例仿真分析。仿真结果表明,平台补贴的头部带动策略对UGC平台优质内容生成系统具有显著的改善效果,平台补贴系数与改善效果具有正相关性且具有可控的合理取值区间。研究结果为UGC平台有效开展内容治理促进平台优质内容生态改善提供了理论依据。

关键词: UGC平台, 优质内容, 微分博弈, 平台补贴, 头部用户