摘要
随着互联网信息技术的迅速发展,网络数据量快速增长,如何在海量数据中找到用户感兴趣的信息并实现个性化推荐是目前重要的研究方向。协同过滤算法作为推荐系统中的经典方法被广泛应用于不同场景,但是仍然存在数据稀疏,以及在计算相似度时不能考虑到所有数据的问题,只能够利用具有共同评分的数据,严重影响了推荐的精确度。针对上述存在的问题,提出了一种融合上下文信息与核密度估计的协同过滤个性化推荐算法。该算法通过对用户和项目各自的上下文信息和已经存在的用户评分数据进行处理,通过核密度估计构建用户和项目的兴趣模型,充分挖掘了用户和项目的兴趣分布,以获得更准确的用户和项目兴趣相似度,降低预测评分误差。在公开的数据集上验证表明,将该算法对比传统的协同过滤算法,有效提高了推荐的精确度。
With the development of Internet information technology and the growth of network data,how to find the information that users are interested in from the massive data and realize personalized recommendation is an important research direction at present.As a classic method in the recommendation system,collaborative filtering algorithm is widely used in different scenes,but it still cannot solve the problem of data sparsity,and in the calculation of similarity,it cannot take all the data into account and can only use the common data,which seriously affects the accuracy of recommendation.Aiming at the problems above,we propose a collaborative filtering recommendation fusing context information and kernel density estimation.The algorithm is based on the user and project their own context information and existing user rating data for processing,based on kernel method respectively to build user and project estimation model,fully taping the interest distribution of user and project,so as obtain more accurate similarity of user and project and reduce the prediction error.The validation on the open data set shows that compared with the traditional collaborative filtering algorithm,the proposed algorithm can effectively improve the accuracy of recommendation.
作者
马瑞新
郭芳清
刘振娇
陈志奎
赵亮
MA Rui-xin;GUO Fang-qing;LIU Zhen-jiao;CHEN Zhi-kui;ZHAO Liang(School of Software Technology,Dalian University of Technology,Dalian 116620,China)
出处
《计算机技术与发展》
2021年第4期34-39,共6页
Computer Technology and Development
基金
国家自然科学基金(61906030)
中央高校基本科研业务费专项资金(DUT20RC(4)009)。
关键词
协同过滤算法
核密度估计
上下文信息
兴趣估计模型
推荐系统
collaborative filtering algorithm
kernel density estimation
context information
interest estimation model
recommendation system
