当前位置: X-MOL 学术Big Data › 论文详情
Our official English website, www.x-mol.net, welcomes your feedback! (Note: you will need to create a separate account there.)
Identifying Influential Nodes Using a Shell-Based Ranking and Filtering Method in Social Networks
Big Data ( IF 4.6 ) Pub Date : 2021-06-16 , DOI: 10.1089/big.2020.0259
Hamid Ahmadi Beni 1 , Asgarali Bouyer 1
Affiliation  

The main goal in the influence maximization problem (IMP) is to find k minimum nodes with the highest influence spread on the social networks. Since IMP is NP-hard and is not possible to obtain the optimum results, it is considered by heuristic algorithms. Many strategies focus on the power of the influence spread of core nodes to find k influential nodes. Most of the core detection-based methods concentrate on nodes in the highest core and often give the same power for all nodes in the best core. However, some other nodes fairly have the potential to select as seed nodes in other less important cores, because these nodes can play an important role in the diffusion of information among the core nodes with other nodes. Given this fact, this article proposes a new shell-based ranking and filtering method, called shell-based ranking and filtering method (SRFM), for selecting influential seeds with the aim to maximize influence in the network. The proposed algorithm initially selects a set of nodes in different shells. Moreover, a set of the candidate nodes are created, and most of the periphery nodes are removed during a pruning approach to reduce the computational overhead. Therefore, the seed nodes are selected from the candidate nodes set using the role of the bridge nodes. Experimental results in both synthetic and real data sets showed that the proposed SRFM algorithm has more acceptable efficiency in the influence spread and runtime than other algorithms.

中文翻译:

在社交网络中使用基于壳的排序和过滤方法识别有影响力的节点

影响最大化问题 (IMP) 的主要目标是找到 k 个在社交网络上传播影响最大的最小节点。由于 IMP 是 NP-hard 并且不可能获得最佳结果,所以它被启发式算法考虑。许多策略侧重于核心节点的影响力传播的威力,以找到 k 个有影响力的节点。大多数基于核心检测的方法都集中在最高核心中的节点,并且通常为最佳核心中的所有节点提供相同的功率。但是,其他一些节点相当有可能在其他不太重要的核心中选择作为种子节点,因为这些节点可以在核心节点与其他节点之间的信息传播中发挥重要作用。鉴于这一事实,本文提出了一种新的基于 shell 的排序和过滤方法,称为基于壳的排序和过滤方法 (SRFM),用于选择有影响力的种子,以最大限度地提高网络中的影响力。所提出的算法最初在不同的壳中选择一组节点。此外,创建了一组候选节点,并且在修剪方法期间移除了大部分外围节点以减少计算开销。因此,种子节点是利用桥节点的角色从候选节点集中选出的。在合成和真实数据集上的实验结果表明,与其他算法相比,所提出的 SRFM 算法在影响传播和运行时间方面具有更可接受的效率。创建一组候选节点,并在修剪方法期间移除大部分外围节点以减少计算开销。因此,种子节点是利用桥节点的角色从候选节点集中选出的。在合成和真实数据集上的实验结果表明,与其他算法相比,所提出的 SRFM 算法在影响传播和运行时间方面具有更可接受的效率。创建一组候选节点,并在修剪方法期间移除大部分外围节点以减少计算开销。因此,种子节点是利用桥节点的角色从候选节点集中选出的。在合成和真实数据集上的实验结果表明,与其他算法相比,所提出的 SRFM 算法在影响传播和运行时间方面具有更可接受的效率。
更新日期:2021-06-18
down
wechat
bug