摘要
在现实生活中,用户对兴趣点的偏好会受到时空场景的影响,用户希望获得匹配当前时间的推荐结果。由此,提出基于用户长短期偏好及时空场景的下一个兴趣点推荐模型。该模型围绕实时兴趣点推荐这一问题,从用户的长短期偏好两方面来挖掘用户的实时兴趣偏好。对于长期偏好,从历史数据中收集与当前时空场景最相关的信息。对于短期偏好,在序列影响的基础上考虑时间推移影响。在公开数据集上的实验结果证明了方法的有效性。
In real life,users’preferences for points of interest are influenced by spatio‑temporal context,and users hope to get recommendation results that match the current time.Therefore,a next point of interest recommendation model based on users’longterm and short‑term preferences and spatio‑temporal context is proposed.This model revolves around the issue of real‑time point of interest recommendation,and explores users’real‑time interest preferences from both long‑term and short‑term preferences.For long‑term preferences,information most relevant to the current spatio‑temporal context is collected from historical data.For shortterm preferences,the influence of time progression is considered on the basis of sequence influence.The experimental results on public datasets prove the effectiveness of the method.
作者
郑宏洲
曾国荪
Zheng Hongzhou;Zeng Guosun(Department of Computer Science and Technology,Tongji University,Shanghai 201804,China)
出处
《现代计算机》
2023年第24期18-25,共8页
Modern Computer
关键词
兴趣点推荐
长短期偏好
时空场景信息
偏好推荐模型
point of interest recommendation
long and short term preference
spatio‑temporal information
preference recommendation model