基于混沌粒子群聚类优化的协同过滤推荐  

Collaborative filtering recommendation based on chaotic particle swarm optimization

作　　者：李建军[1,2] 付佳 杨玉[1,2] 汪校铃[1,2] 荣欣 LI Jian-jun;FU Jia;YANG Yu;WANG Xiao-ling;RONG Xin(School of Computer and Information Engineering,Harbin University of Commerce,Harbin 150028,China;Heilongjiang Provincial Key Laboratory of Electronic Commerce and Information Processing,Harbin University of Commerce,Harbin 150028,China)

机构地区：[1]哈尔滨商业大学计算机与信息工程学院,黑龙江哈尔滨150028 [2]哈尔滨商业大学,黑龙江省电子商务与信息处理重点实验室,黑龙江哈尔滨150028

出　　处：《计算机工程与设计》2021年第8期2173-2179,共7页Computer Engineering and Design

基　　金：国家自然科学基金项目(60975071);黑龙江省新型智库研究基金项目(18ZK015);黑龙江省哲学社会科学研究规划基金项目(17GLE298、16EDE16);哈尔滨商业大学校级课题基金项目(18XN065);哈尔滨商业大学博士科研启动基金项目(2019DS029)。

摘　　要：为解决复杂的网络信息无法对用户进行精准推荐的情况,改进传统协同过滤算法,将混沌粒子群算法与协同过滤算法融合使用。在传统粒子群算法中加入混沌扰动并随着迭代调整惯性权重,对用户进行聚类优化。获取目标用户之后,通过判断目标用户属于哪个聚类,在该聚类内部进行协同过滤计算。通过与其它算法之间的对比实验,验证了基于混沌粒子群聚类优化的协同过滤推荐算法相较其它算法具有更低的平均绝对误差和更高的准确率。To solve the situation that complex network information cannot accurately recommend users,the traditional collaborative filtering algorithm was improved,and the chaotic particle swarm optimization algorithm and the collaborative filtering algorithm were used together.In the traditional particle swarm optimization algorithm,chaotic disturbances were added and the inertia weight was adjusted with iteration to optimize the user clustering.After acquiring the target user,the cluster to which the target user belongs was determined,and collaborative filtering calculation was performed within the cluster.Through comparison experiments with other algorithms,it is verified that the collaborative filtering recommendation algorithm based on chaotic particle swarm optimization has lower average absolute error and higher accuracy than other algorithms.

关 键 词：混沌系统 粒子群算法 聚类优化 协同过滤算法 推荐 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TP301.6[自动化与计算机技术—控制科学与工程]