基于局部联合矩阵转移的PageRank图个性化分析

摘要

为提高图分析中PageRank模型的计算效率,提出基于联合局部敏感转移矩阵的PageRank模型改进形式.设计联合局部及边界的局部个体转移矩阵,将计算限定在图局部,降低节点参与转移矩阵的计算数量,提高计算效率,为保证算法精度,给出该方式的补偿矩阵;在矩阵计算过程中,考虑可重复利用矩阵计算的再利用问题,降低计算的重复性,进一步提高计算效率,给出所设计算法的计算复杂度分析.在标准数据集中的仿真测试结果表明,与FRWR和GMES等算法相比,所提算法具有更高的计算精度和效率.%To improve the computational efficiency of PageRank model,a PageRank model based on joint local sensitive transfer matrix was proposed.The local individual transfer matrix was designed,which limited the calculation to the local,and reduced the number of nodes participating in the transfer matrix,to improve the calculation efficiency.The compensation matrix was given to guarantee algorithm accuracy.In the process of matrix calculation,the reuse problem was considered,which reduced the computational complexity and improved the computational efficiency,and also the computational complexity analysis of the algorithm was given.Results of the simulation in the standard data set show that the proposed algorithm has higher accuracy and efficiency compared with FRWR and GMES.