Data-centric Decision Supports for Large-scale Social Network Management

以数据为中心的大规模社交网络管理决策支持

• 批准号: RGPIN-2017-05667
• 负责人: Li, Zhepeng
• 金额: $ 1.46万
• 依托单位: York University
• 依托单位国家: 加拿大
• 项目类别: Discovery Grants Program - Individual
• 财政年份: 2020
• 资助国家: 加拿大
• 起止时间: 2020-01-01 至 2021-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=711352
• 关键词: Data; centric; Decision; Supports; Large

Growing interest in the value of digital information, commonly known as data, creates opportunities to re-invent managerial machinery and lead to the changes in decision making. A much desired means to gain value from data stems from decision support (DS), which provides information that supports users and organizations in decision-making activities. To embrace the advent of big data, decision support needs to address challenges in discovering actionable knowledge when the problem scales up and uncertainty arises substantially. Data-centric approaches are advocated to cope with these issues. This proposal is intended to investigate a theme of decision support problems that caters to the needs of effective management in large-scale social networks. The formulated problems are to be addressed by data-centric approaches integrating machine learning, optimization, and heuristic methods. In the long run, it aims at achieving a comprehensive DS toolbox for effective social network management and governance. As for the short-term goal, it is to focus on the predictive stage that addresses actionable knowledge for decision support problems in social network formation, network-based propagation, and social recommendation. In Canada, a large portion of the population is engaged with online social networking services. At the same time, the business of social networks is estimated to grow considerably. The proposed research will generate benefits in terms of both commercial and societal interests. Methods that are developed in the proposed research can be applied in a wide range of domains, in that the large network can conveniently adapt to account for financial interactions, world-wide web, genetic graph, and terrorism relationships. The focal research would thus accomplish external impacts by addressing related problems, such as credit/risk evaluations, WWW navigability, identification of hidden protein interactions, and terrorism/riots surveillance/containment, which range from finance, IT, biology to sociology.

对数字信息（通常称为数据）价值的兴趣日益增长，为重新发明管理机制创造了机会，并导致决策的变化。从数据中获取价值的一种非常理想的方法来自决策支持（DS），它提供支持用户和组织决策活动的信息。为了迎接大数据的到来，决策支持需要在问题扩大和不确定性大幅增加时解决发现可操作知识的挑战。提倡以数据为中心的方法来科普这些问题。 该建议旨在研究满足大规模社交网络有效管理需求的决策支持问题。制定的问题将通过以数据为中心的方法来解决，这些方法集成了机器学习，优化和启发式方法。从长远来看，它旨在实现一个全面的DS工具箱，用于有效的社交网络管理和治理。至于短期目标，它是专注于预测阶段，解决社会网络形成，基于网络的传播和社会推荐中的决策支持问题的可操作知识。 在加拿大，很大一部分人口从事在线社交网络服务。与此同时，社交网络的业务预计将大幅增长。拟议的研究将在商业和社会利益方面产生效益。在拟议的研究中开发的方法可以应用于广泛的领域，因为大型网络可以方便地适应金融互动，万维网，遗传图和恐怖主义关系。因此，重点研究将通过解决相关问题来实现外部影响，例如信用/风险评估，WWW导航性，隐藏蛋白质相互作用的识别，以及恐怖主义/骚乱监视/遏制，范围从金融，IT，生物学到社会学。