摘要
大数据背景下,数据挖掘及信息推荐越来越普遍,但目前推荐系统主要集中在单一数据来源,忽视源数据的多样性。文中根据用户历史或当前操作对推荐数据的影响定义数据集的可用性,考虑数据间可用度的相互影响、用户对不同源数据集的偏好程度及偏好的周期变化,结合关联规则推荐和协同推荐算法的优点,提出多样性数据推荐算法。
Under the background of big data,data mining and information recommendation become more and more common,but at present recommendation system mainly concentrated in a single data source,without considering the diversity of the source data. This paper,according to the influence of the user＇s history or current operation of the recommended data,defines the availability of the data set. It proposes a diversified data recommendation algorithm by considering the influence of availability between the different data,the user preferences of different source data and the periodic change of user preferences.
出处
《信息技术》
2017年第8期140-144,148,共6页
Information Technology
关键词
多样性
推荐算法
可用性
可信度
diversity
recommendation algorithm
availability
credibility
