融合局部与全局紧密度的符号网络链接预测算法

鉴于大多数符号网络预测算法仅能对已有链接缺失的符号进行预测,无法实现未知的链接及其符号预测,提出一种融合局部与全局结构特征定义节点间相似性的符号网络链接预测算法.基于结构平衡理论,利用连接两节点的步长为2和3的路径信息分别定义局部和全局链接紧密度,有效融合两者得到两节点的总相似度,其绝对值度量了链接建立的可能性,其符号即为链接的符号预测结果.在多个经典的符号网络数据集上对算法的有效性和正确性进行了验证,并与符号网络中有代表性的预测算法进行了准确率以及推荐链接的对比分析.结果显示,所提算法在链接预测与符号预测两方面均达到了较好的预测性能.

Link prediction in signed networks based on local and global tightness

Most prediction algorithms in signed networks can only predict the missing sign of existing links, and they couldn''t achieve the prediction of unknown links and their signs. In view of this, this paper proposed a link prediction algorithm for signed networks, which integrated local and global structural features to define the similarity. Based on the structural balance theory, it defined the local tightness and global tightness respectively by using the structure information of paths with a step size of 2 and 3. Then it could get the total similarity of the two nodes, where the absolute value measured the possibility of establishing a link, and its sign was the sign prediction result of the link. It verified the effectiveness and correctness of this algorithm on several classic datasets. And it also did comparison and analysis of the accuracy and recommended links with other representative prediction algorithms in signed networks. Experimental results show that the proposed algorithm performs well in both link prediction and sign prediction.