北京邮电大学学报

  • EI核心期刊

北京邮电大学学报 ›› 2018, Vol. 41 ›› Issue (6): 115-122.doi: 10.13190/j.jbupt.2018-028

• 研究报告 • 上一篇    下一篇

基于粒子群优化算法的协同过滤推荐并行化研究

游思晴, 周丽, 赵东杰, 薛菲   

  1. 北京物资学院 信息学院, 北京 101149
  • 收稿日期:2018-01-26 出版日期:2018-12-28 发布日期:2018-12-24
  • 作者简介:游思晴(1982-),女,讲师,E-mail:93028603@qq.com.
  • 基金资助:
    国家自然科学基金项目(71501015);北京市智能物流系统协同创新中心开放课题

Research on Parallelization of Collaborative Filtering Recommendation Algorithm Based on Particle Swarm Optimization

YOU Si-qing, ZHOU Li, ZHAO Dong-jie, XUE Fei   

  1. School of Information, Beijing Wuzi University, Beijing 101149, China
  • Received:2018-01-26 Online:2018-12-28 Published:2018-12-24

摘要: 针对常用协同过滤推荐算法存在计算性能瓶颈的问题,提出了在Spark上并行化实现协同过滤推荐算法RLPSO_KM_CF.首先,通过具备反向学习和局部学习能力的粒子群优化(RLPSO)算法寻找粒子群最优解,输出优化后的聚类中心;然后,运用RLPSO_KM算法对用户信息进行聚类;最后,将传统协同过滤推荐算法与RLPSO_KM聚类结合,从而对目标用户进行有效推荐.实验结果显示,RLPSO_KM_CF算法在推荐准确度方面有显著提高,具有较高的加速比,稳定性也得到了一定提升.

关键词: 协同过滤推荐算法, RLPSO算法, K-means算法, Spark

Abstract: In order to solve the computational performance bottleneck of the commonly used collaborative filtering recommendation algorithm, a parallel collaborative filtering recommendation algorithm RLPSO_KM_CF on Spark is proposed. Firstly, the reverse-learning and local-learning particle swarm optimization (RLPSO) algorithm is used to find the optimal solution of the particle swarm and the output clustering center is optimized. Then, the RLPSO_KM algorithm is used to cluster the user information. Finally, the traditional cooperative filtering recommendation algorithm is combined with the RLPSO_KM cluster to effectively recommend the target user. The experimental results show that the improved algorithm has a significant improvement in the recommended accuracy, and has a higher speedup and stability.

Key words: collaborative filtering recommendation algorithm, RLPSO algorithm, K-means algorithm, Spark

中图分类号: