基于情景化用户偏好的信息推荐研究  

Research on Information Recommendation Based on Contextualized User Preference

作　　者：李鸿勋 杨君[2] LI Hongxun;YANG Jun(Department of Computer Science,Hong Kong Baptist University,Hongkong 999077,China;School of Management,Guangdong University of Technology,Guangzhou 510006,China)

机构地区：[1]香港浸会大学理学院,中国香港999077 [2]广东工业大学管理学院,广东广州510006

出　　处：《现代信息科技》2021年第22期1-3,7,共4页Modern Information Technology

基　　金：广东省哲学社会科学规划项目(GD17CTS03);广东省自然科学基金项目(2018A030313933)。

摘　　要：以支付宝为研究对象探讨了情景化用户偏好对信息推荐的影响,提出了基于情景化用户偏好的信息推荐流程,改进了已有的推荐算法。将传统的协同过滤算法中由项目评分预测项目评分的模式调整为以用户收支模式预测用户的功能使用情况的模式。在推荐结果输出阶段,通过输出情景化将时间这一情景因素融入推荐结果的输出中,在合适的时间段向用户进行对应功能的推荐来提高用户对推荐结果的接受程度。Taking Alipay as the research object, the influence of contextualized user preferences on information recommendation is discussed, and the information recommendation process based on contextualized user preferences is proposed, and the existing recommendation algorithm is improved. In the traditional collaborative filtering algorithm, the mode of predicting project score from project score is adjusted to the mode of predicting user function usage based on user revenue and expenditure mode. In the recommendation result output stage, the context factor of time is integrated into the output of the recommendation result through the output of context,and the corresponding function is recommended to the user in an appropriate time period to improve the user’s acceptance of the recommendation results.

关 键 词：情景 情景感知 用户偏好 信息推荐 

分 类 号：TP391[自动化与计算机技术—计算机应用技术] G203[自动化与计算机技术—计算机科学与技术]