摘要
针对现有序列推荐算法时间信息利用不充分的问题,提出一种多时间嵌入模式,用动态融合策略缓解现有序列推荐算法长期偏好建模不充分的问题.多时间嵌入模式能同时建模用户物品交互的绝对时间信息和相对时间信息,充分捕获用户物品交互关于时间的多种规律.动态融合网络根据用户意图动态融合用户的长期偏好和近期偏好,精准刻画用户兴趣,提升推荐结果的多样性.在公共数据集MovieLens-1M和Amazon-Beauty上将该长期记忆增强的时间感知序列推荐算法与现有算法进行对比,结果表明,该算法在评价指标HR@N和NDCG@N上均较对比方法有提高.实验结果表明,长期记忆增强的时间感知序列推荐算法较其他对比序列推荐算法准确率有一定提高.
Aiming at the problem that the existing sequence recommendation algorithms did not make full use of time information,we proposed a multi-time embedding mode,which used dynamic fusion strategy to alleviate the problem of insufficient long-term preference modeling in existing sequence recommendation algorithms.The multi-time embedding mode could simultaneously model the absolute time information and relative time information of user-item interaction,and fully capture various rules of user-item interaction about time.The dynamic fusion network dynamically integrated the users’long-term preference and recent preference according to the users’intentions,accurately depicted the users’interest,and improved the diversity of recommendation results.The proposed time embedding sequence recommendation algorithm based on long-term memory enhancement was compared with the existing algorithms on the public datasets MovieLens-1M and Amazon-Beauty.The results show that the proposed algorithm is better than the comparison method in the evaluation indicators HR@N and NDCG@N.Experimental results show that the accuracy of the time embedding sequence recommendation algorithm based on long-term memory enhancement is higher than that of other comparison sequence recommendation algorithms.
作者
陈继伟
汪海涛
朱兴翔
姜瑛
陈星
CHEN Jiwei;WANG Haitao;ZHU Xingxiang;JIANG Ying;CHEN Xing(Faculty of Information Engineering and Automation,Kunming University of Science and Technology,Kunming 650504,China;SPIC Yunnan International Power Investment Co.,Ltd.,Kunming 650100,China)
出处
《吉林大学学报（理学版）》
CAS
北大核心
2022年第4期919-928,共10页
Journal of Jilin University:Science Edition
基金
国家自然科学基金(批准号:61462049).
关键词
深度学习
推荐算法
注意力机制
时间嵌入
长短期兴趣
deep learning
recommendation algorithm
attention mechanism
time embedding
long-term and short-term interest
