摘要
文中从用户行为及时间效应两方面出发,研究出了基于用户行为数据的个性化推荐算法。首先,研究了基于用户行为数据时间效应的推荐算法,分析根据线性时间加权推荐算法中的问题。其次,由于用户兴趣并不只是单纯的根据时间的线性变化,其主要是根据时间段发生变化,故文中分析了与时间窗口技术结合的推荐算法。最后,结合时间效应及并行化推荐算法进行改进,实现了个性化推荐算法。结果表示,此推荐算法能够对推荐结果进行优化,使算法精准性得到有效的提高。
This paper studies the personalized recommendation algorithm based on user behavior data from two aspects of user behavior and time effect. Firstly,the recommendation algorithm based on the time effect of user behavior data is studied,and the problem in the linear time-weighted recommendation algorithm is analyzed. Secondly,since user interest is not simply a linear change according to time,it mainly changes according to the time period. Therefore,this paper analyzes the recommendation algorithm combined with time window technology. Finally,combined with the time effect and parallel recommendation algorithm,the personalized recommendation algorithm is implemented. The result shows that this recommendation algorithm can optimize the recommendation results and effectively improve the accuracy of the algorithm.
作者
皇甫汉聪
肖招娣
HUANGFU Han-cong;XIAO Zhao-di(Foshan Power Supply Bureau,Guangdong Power Grid Company,Foshan 528000,China)
出处
《电子设计工程》
2019年第7期38-41,46,共5页
Electronic Design Engineering
关键词
用户行为
数据分析
个性化推荐
算法研究
user behavior
data analysis
personalized recommendation
algorithm research
