多维社会网络中链路预测方法研究

LINK PREDICTION METHODS RESEARCH IN MULTI-DIMENSIONAL SOCIAL NETWORKS

多维社会网络中链路预测是指通过已知的网络节点以及多维社会网络结构等信息预测多维社会网络中尚未产生连边的两个节点之间产生链接的可能性。在研究有关链路预测的相关技术后,基于节点之间的相似性,提出了一种多维社会网络中的链路预测算法。算法中考虑时间因素和权重因素对于链路预测的影响,同时将多维网络中的相关技术应用到链路预测的过程中。实验结果表明,该算法能够使多维网络中链路预测有更好的效果。

Link prediction in multi-dimensional social networks refers to estimating,on the basis of known network nodes as well as multi-dimensional social network structure and other information,the likelihood of the generation of links between two nodes that are temporarily unlinked in multi-dimensional social networks.After examining relevant link prediction technologies,based on similarities among nodes,a link prediction algorithm in multi-dimensional social networks is presented.In the algorithm,time and weight factors are taken into consideration for their influence on link prediction,while relevant technologies in multi-dimensional networks are applied to the link prediction process.Experimental results show that the algorithm can improve the link prediction performance in multi-dimensional networks.