摘要
大数据时代数据量呈爆发式增长,为帮助人们在海量数据中获取自己所感兴趣的信息,推荐系统应运而生。协同过滤在推荐系统中应用广泛,针对传统协同过滤推荐算法数据稀疏、推荐精度较低,不能及时反映用户的兴趣度变化以及时效性不足等缺点,提出了一种融合时间衰减函数的改进协同过滤推荐算法。此算法在传统协同过滤算法的基础上综合考虑了时间因素的影响,用户兴趣会随着时间而变化,用户在短时间内感兴趣的物品具有更高的相似性,参考人类记忆遗忘特性,拟合人类记忆遗忘曲线得到时间衰减函数作为权重因子,在计算相似度和用户偏好程度时同时融入时间衰减函数对算法进行约束,提高短时间内物品相似度和用户兴趣度的权重,实现短期和长期兴趣度融合。实验结果表明,改进后的方法能在一定程度上提高传统推荐算法的精确率和召回率,验证了时间衰减函数的有效性。
In the era of big data,the amount of data grows explosively.To help people get the information they are interested in from the mass data,the recommendation system emerges at the historic moment.Collaborative filtering is widely used in recommendation systems.Aiming at the shortcomings of traditional collaborative filtering recommendation algorithms,such as sparse data,low precision,inability to timely reflect the changes of users’ interest and lack of timeliness,we propose an improved collaborative filtering recommendation algorithm based on fusion of time decay function.Based on the traditional collaborative filtering algorithm,the proposed algorithm comprehensively takes the influence of time factor into account.Users’ interest will change with time,and the items that users are interested in in a short period of time have higher similarity.By referring to the law of human memory forgetting and fitting the curve of human memory forgetting,the time decay function is obtained as the weight factor.In the calculation of similarity and user preference,the algorithm is constrained by the time decay function,so as to improve the weight of similarity and user interest in a short period of time,and realize the combination of short-term and long-term interest.Experiment shows that the improved method can improve the precision and recall rate of the traditional recommendation algorithm to a certain extent,which verifies the validity of the time decay function.
作者
殷佳莉
江智威
杨毅
刘培培
YIN Jia-li;JIANG Zhi-wei;YANG Yi;LIU Pei-pei(School of Information Science and Technology(School of Cyber Security,Oxford Brookes College),Chengdu University of Technology,Chengdu 610059,China)
出处
《计算机技术与发展》
2022年第4期170-174,共5页
Computer Technology and Development
基金
国家自然科学基金青年基金(11905020)。
关键词
推荐算法
协同过滤
人类记忆遗忘特性
时间衰减函数
兴趣偏好
recommendation algorithm
collaborative filtering
law of forgetting human memory
time decay function
interest preference
