A Hybrid Personalized Recommender System for Content Suggestion

用于内容建议的混合个性化推荐系统

• 批准号: 508327-2017
• 负责人: Ding, Chen
• 金额: $ 1.82万
• 依托单位: Ryerson University
• 依托单位国家: 加拿大
• 项目类别: Engage Grants Program
• 财政年份: 2017
• 资助国家: 加拿大
• 起止时间: 2017-01-01 至 2018-12-31
• 项目状态: 已结题
• 来源: https://www.nserc-crsng.gc.ca/ase-oro/Details-Detailles_eng.asp?id=640113
• 关键词: Hybrid; Personalized; Recommender; System; Content

The proposed research is in the area of recommender systems and predictive analytics, and should have animpact in Canadian ICT (Information and Communications Technology) sector. PostBeyond is aToronto-based company providing employee communications solutions to primarily large enterprises. UsingPostBeyond's product, company employees can have easy access to company-related information through acentral content hub, with the option to share the information externally across their social networks. Thecurrent solution is one size fit all. Sometimes it could be hard for individual employees to find content they areparticularly interested in and would like to share with others. To solve this problem, in this project, we proposeto build a personalized recommender system that could suggest content for employees to access, consume andshare based on their individual preferences. A few research challenges we must tackle include: 1) the lack ofhistorical data on users' content access and sharing patterns; 2) the need to define content profile that is bothgeneric and extensible; 3) the need to accommodate users' dynamic behavior into the system; 4) the difficultyof effectively integrating group recommendation features into the personalized recommendation system. Tosuccessfully build the proposed recommender system, the following three tasks are identified: 1) developing alogging tool to save users' content access and social sharing records; 2) applying log mining techniques to minethe user patterns, behaviors and observe the general content consumption or sharing trends; 3) designing andimplementing a hybrid recommender system to do personalized content suggestion and prediction. Theproposed project could benefit PostBeyond by making their current communications solutions more efficientand customized towards each individual employee's behavior, and thus a more attractive solution to theirpotential customers. It also drives economic benefits for both PostBeyond and the many Canadian enterprisesthat use their solutions.

拟议的研究是在推荐系统和预测分析领域，应该会对加拿大的信息和通信技术(ICT)部门产生影响。PostBeyond是一家总部位于多伦多的公司，主要为大型企业提供员工通信解决方案。使用PostBeyond的产品，公司员工可以通过中央内容中心轻松访问公司相关信息，并可以选择通过他们的社交网络在外部共享信息。目前的解决方案是一刀切。有时，个别员工可能很难找到他们特别感兴趣并希望与他人分享的内容。为了解决这一问题，在本项目中，我们提出了构建一个个性化推荐系统，该系统可以根据员工的个人偏好推荐内容供他们访问、消费和分享。我们必须解决的一些研究挑战包括：1)缺乏关于用户内容访问和共享模式的历史数据；2)需要定义既通用又可扩展的内容配置文件；3)需要在系统中容纳用户的动态行为；4)将组推荐特征有效地集成到个性化推荐系统中。为了成功地构建推荐系统，确定了以下三个任务：1)开发日志工具来保存用户的内容访问和社交分享记录；2)应用日志挖掘技术来挖掘用户的模式、行为和观察一般的内容消费或分享趋势；3)设计和实现一个混合式推荐系统来进行个性化的内容建议和预测。拟议中的项目可以使PostBeyond受益，因为它使他们目前的通信解决方案更有效率，并针对每个员工的行为进行定制，从而成为对潜在客户更具吸引力的解决方案。它还为PostBeyond和使用其解决方案的许多加拿大企业带来了经济效益。