基于社区划分的多线程潜在好友推荐算法

发布时间：2019-04-30 19:46

【摘要】：针对基于社区划分的潜在好友推荐算法FRCD运行速度慢的问题,提出了一种基于社区划分的多线程潜在好友推荐算法MTFRCD。该算法在网络拓扑图上利用多线程技术寻找核心关系子网,以核心关系子网作为标签种子节点,使用多线程并发传播标签来发现网络拓扑图上的社区结构,利用社区发现结果在社区内部推荐潜在好友。人工网络的实验结果表明,MTFRCD相比于传统的FRCD,在性能近似的前提下具有明显的速度增长。因此,将该算法应用于真实社交网络(学者网)平台的潜在好友挖掘和推荐,根据推荐结果的评测,验证了算法具有良好的推荐效果。
[Abstract]:In order to solve the problem of slow running speed of potential friend recommendation algorithm FRCD based on community partition, a multi-thread potential friend recommendation algorithm MTFRCD. based on community partition is proposed. The algorithm uses multi-thread technology to find the core relation subnet on the network topology diagram, uses the core relation subnet as the label seed node, and uses the multi-thread concurrent propagation label to discover the community structure on the network topology diagram. Using community discovery results to recommend potential friends within the community. The experimental results of artificial network show that compared with the traditional FRCD, MTFRCD has a significant increase in performance under the premise of approximate performance. Therefore, the algorithm is applied to mining and recommending potential friends on the real social network (Scholar Network) platform. Based on the evaluation of the recommendation results, the proposed algorithm is proved to have a good recommendation effect.
【作者单位】：华南师范大学计算机学院;广东技术师范学院计算机科学学院;
【基金】：国家“863”计划资助项目(2013AA01A212) 广州市科技计划资助项目(2014J4300033) 2014年广东省重大科技专项资助项目(2014B010116002) 广东省自然科学基金资助项目(2015A030310509) 国家自然科学基金青年科学基金资助项目(61502180) 广东省科技计划资助项目(2015B010109003,2013B0908000024)
【分类号】：TP393.09;TP391.3

【相似文献】