一种适用于“电力数据口袋书”业务的多维度协同过滤推荐算法设计

Design of a Multi-Dimensional Collaborative Filtering Recommendation Algorithm Suitable for"Power Data Pocket Book"Service

本文提出了一种使用知识图谱叠加喜好的多维度智能推荐算法,详细地讨论了协同推荐算法,并对其增量更新、并行计算、加权版等方面的内容进行了分析,此算法推荐准确度高、易于实现,很适用于消息推荐服务。提出协同推荐算法用在消息推送服务中,对用户数据作分析,筛选出推送消息适合的目标用户,从而提高推送质量;设计并实现了服务于“电力数据口袋书”的消息智能推送服务系统,其能智能对推送用户作筛选,提高推送质量。系统向应用开发者提供友好的API,能在高并发环境下稳定运行,支持设备的别名、标签设置,支持消息优先级、定时推送等功能。

In this paper,a multi-dimensional intelligent recommendation algorithm using knowledge graph superposition preferences is proposed.The collaborative recommendation algorithm is discussed in detail,and its incremental update,parallel computing,and weighted version are analyzed.This algorithm has high recommendation accuracy,is easy to implement,and is very suitable for message recommendation services.The collaborative recommendation algorithm is proposed to be used in the message push service,which analyzes the user data and filters out the target users suitable for the push message,so as to improve the push quality.A message intelligent push service system for"Power Data Pocket Book"is designed and implemented,which can intelligently screen the push users and improve the push quality.The system provides application developers with friendly API,which can run stably in a high concurrency environment,support device alias and label settings,support message priority,regular push and other functions.