Influence maximisation is the problem of selecting a specific number of nodes which can maximise the influence spread of social networks. For its significant practical applications, the influence maximisation problem has been widely used in many fields, such as network marketing and rumour control. However, most existing algorithms tend to select accuracy or efficiency to optimise, which leads to their poor performance. Therefore, a Three-step Fast Selection algorithm for Influence Maximisation (TsFSIM) is proposed in this paper. Firstly, a new method to evaluate nodes' influence spread is proposed, called Influence Estimation Value. Influence Estimation Value (IEV) combines the node's and its neighbours' degree to estimate its influence. This can improve the efficiency of our algorithm. Afterwards, based on IEV, a three-stage filtering strategy is proposed. This strategy can improve the accuracy of our algorithm greatly. Finally, experimental results on seven real-world networks show that the proposed method is more accurate than other methods while keeping competitive efficiency.

影响力最大化是选择特定数量的节点来最大化社交网络的影响力传播的问题。由于其重要的实际应用，影响力最大化问题已被广泛应用于许多领域，例如网络营销和谣言控制。然而，大多数现有算法倾向于选择精度或效率进行优化，这导致其性能不佳。因此，本文提出了一种影响最大化的三步快速选择算法（TsFSIM）。首先，提出了一种评估节点影响力传播的新方法，称为影响估计值。影响估计值 (IEV) 结合节点及其邻居的程度来估计其影响。这可以提高我们算法的效率。之后，基于IEV，提出了一种三级过滤策略。这种策略可以大大提高我们算法的准确性。最后，在七个真实世界网络上的实验结果表明，在保持竞争效率的同时，所提出的方法比其他方法更准确。