社交网络用户反馈数据的个性化推荐算法仿真  

Simulation of Personalized Recommendation Algorithm for User Feedback Data of Social Networks

作　　者：冯必波[1] 张伶俐[2] 尹静[1] FENG Bi-bo;ZHANG Ling-li;YIN Jing(School of Computer Science and Engineering,Chongqing University of Technology,Chongqing 400054,China;Chongqing University of Technology,Chongqing 400054,China)

机构地区：[1]重庆理工大学计算机科学与工程学院,重庆400054 [2]重庆理工大学,重庆400054

出　　处：《计算机仿真》2024年第11期375-379,共5页Computer Simulation

摘　　要：社交网络通常涉及大量的用户和内容,使得对所有用户和所有内容进行个性化推荐的计算复杂度非常高,导致推荐系统难以捕捉到用户的偏好和兴趣。因此,提出一种社交网络用户反馈数据的个性化推荐算法。通过用户时间和空间范围的相似度计算方法,结合用户之间的信任关系,准确地捕捉目标用户的个性化信息。采用全局相似度计算方法,结合用户邻居和项目邻居的相似度计算,建立全局相似度计算矩阵分解模型,采用聚类算法按用户聚类特征个性化推荐,通过构建兴趣图谱和计算用户与聚类主题之间的参与关系,计算个性化向量的相互距离和差异性指数,实现精准的个性化推荐。实验结果表明,所提方法在困惑度和平均余弦相似性指标上表现好,说明上述方法在同等的用户反馈数据个性化兴趣推荐条件下能够提供精准和与用户兴趣相似的推荐结果。Generally,social networks involve large numbers of users and content,so the computational complexity of personalized recommendations for all users and content is very high,and it is difficult for recommendation systems to capture users'preferences and interests.Therefore,this paper proposed a personalized recommendation algorithm for user feedback data in social network.Firstly,we calculated the similarity between time and space,and combined it with the trust relationship between users to accurately capture the personalized information of the target user.Secondly,we combined the global similarity method with the similarity of user neighbors and item neighbors to build a matrix decomposition model for global similarity calculation.Then,we adopted the clustering algorithm to provide personalized recommendations based on user clustering characteristics.Finally,we constructed an interest graph and calculated the participation relationship between users and clustering topics.Meanwhile,we calculated the mutual distance and difference index between personalized vectors.Thus,we achieved accurate personalized recommendations.Experimental results show that the confusion and average cosine similarity indicators are satisfactory,indicating that the method can provide precise and similar recommendation results that are similar to user interests under the same condition of personalized interest recommendation.

关 键 词：社交网络 用户反馈数据 个性化推荐 活动区域 相似度计算 

分 类 号：TP301[自动化与计算机技术—计算机系统结构]