Finding a number of nodes that are able to maximize the spread of influence through the social network and are called influence maximization has numerous applications in marketing. One such application is to find influential members for promoting a product across a large network. Even though numerous algorithms have been proposed, challenges such as scalability, time constraints, and low accuracy have motivated the researchers for better solutions. Some of the newly proposed algorithms are scalable, but fail to provide adequate accuracy. On the other hand, some greedy algorithms provide a good level of accuracy but are very time consuming for large networks. In this paper, an algorithm is proposed called FAIMCS that can quickly find influential nodes across large networks with high accuracy. FAIMCS, reduces computational overhead considerably by eliminating major portions of the social network graph which have little influence. FAIMCS uses community detection algorithm to determine each community's quota of influential nodes based on the structure of that community. Finally, it obtains influential nodes from the candidate nodes. Experiment results show FAIMCS is faster than current algorithms and provides a high level of accuracy for large social networks.

中文翻译：

---

FAIMCS：一种基于社区结构的社交网络中快速准确的影响力最大化算法

寻找能够通过社交网络最大限度地传播影响力并被称为影响力最大化的节点在营销中有许多应用。一种这样的应用是寻找有影响力的成员来在大型网络上推广产品。尽管已经提出了许多算法，但可扩展性、时间限制和低准确性等挑战促使研究人员寻求更好的解决方案。一些新提出的算法是可扩展的，但不能提供足够的准确性。另一方面，一些贪心算法提供了很好的准确性，但对于大型网络来说非常耗时。在本文中，提出了一种称为 FAIMCS 的算法，该算法可以在大型网络中以高精度快速找到有影响力的节点。FAIMCS, 通过消除影响不大的社交网络图的主要部分，大大减少了计算开销。FAIMCS 使用社区检测算法根据社区的结构确定每个社区的影响节点配额。最后，从候选节点中获取有影响力的节点。实验结果表明 FAIMCS 比当前算法更快，并且为大型社交网络提供了高水平的准确性。