面向社交服务平台的用户复杂消费行为建模与平台演化研究

Social Networking Service(SNS) websites provide a social platform for users to share their consumption preferences over products. Accurately mining and understanding users’ behaviors in these platforms has significant research and marketing values, thus it has become a research hotspot in recent years. However, existing research usually focused on one specific aspect of users’ consumption behavior and could not model users’ complex consumption decision process. Furthermore, the current works seldom considered the interplay between users’ two kinds of behaviors (i.e., user-user link behavior and user-product consumption behavior) in SNSs, thus could not describe the evolution of the platform. To this end, in this proposal, we propose to address the key problems in modeling users’ complex consumption behavior and platform evolution in SNSs. Firstly, to better explain users’ dynamic consumption interest over time, the competition among products is introduced in user interest modeling techniques. Then, by leveraging the social influence theory in social networks, we propose an interest-aware influence model that considers each user’s unique consumption interest in product diffusion and consumption process. The above two models uncover the complex consumption decisions of users. After that, by borrowing the interplay of users’ two kinds of behaviors from social theories, we then design a joint-evolving method to modeling the evolution of users’ two kinds of behaviors in SNSs. Finally, we shift from passive user interest modeling to active user interest guidance, and design a marketing strategy by recommending social links and selecting seed users to guide and maximize the further consumption in SNSs. This proposal presents a deep and extensive framework to understand users’ behaviors in SNSs, which will advance both the theoretical and practical values in user behavior modeling area.

社交服务网站是在Web2.0背景下形成的社交化用户—产品消费分享平台。挖掘用户在该类平台下的行为蕴含巨大的科学和市场价值，因此已成为近年来的一个热点研究方向。然而，传统研究工作通常简化建模用户消费行为，难以实现用户复杂消费行为理解。而且，当前研究较少考虑社交服务平台中用户两类行为（社交与消费）的关联性，揭示该类平台的演化机理。针对上述问题，本课题以社交服务网站中用户两类行为为研究对象，系统性的开展用户复杂消费行为建模与社交服务平台演化研究。具体内容包括：（1）针对用户消费兴趣的时序易变性，研究产品竞争关系下的用户消费兴趣模型；（2）结合社交网络中的传播特性，设计融合用户消费兴趣的产品传播方法；（3）针对平台中用户两类行为的时序相关性，设计平台演化及预测模型；（4）提出面向平台商家的营销策略，主动引导平台后续演化发展。本课题将有力推动社交服务网站中用户复杂消费行为理解的技术研究及应用拓展。

对社交服务平台中的用户行为(用户-产品消费行为，用户-用户社交行为)进行建模分析及研究，对理解用户复杂兴趣爱好及平台演化机理有着重要的理论意义与应用价值。然而，当前研究主要简化用户消费行为，难以利用复杂情境信息进行用户行为建模及平台全局演化研究。针对上述问题，本项目以社交平台中的用户为中心，以用户的两类行为为研究对象，系统性的开展用户复杂行为理解与平台研究。具体研究内容包括：基于社交影响力的个性化推荐方法，社交情境感知的用户及平台演化研究，基于多源异构信息的用户复杂消费行为建模，及推荐系统基础方法研究。在本项目的支持下，已发表学术论文22篇，包括以第一或通讯作者IEEE/ACM Trans.系列论文5篇和CCF推荐的A类会议论文4篇，并与科大讯飞、微软、腾讯等公司建立了密切合作关系。本项目通过探索社交平台复杂用户消费行为机理，提出了社交平台中的用户行为建模及平台演化模型，拓展了社交平台的应用领域，提升了社交网络分析与个性化推荐的应用效果。