Accurate community detection plays a crucial role in understanding the functions and evolution of complex networks. Traditional community detection methods often struggle to achieve satisfactory results when confronted with challenges such as high overlap, blurred boundaries, and structural sparsity. To address this, this paper proposes a new overlapping community detection method based on Seeded Similarity Fusion and Boundary Node Walk(SSFBW). The algorithm operates as follows: we design a novel seed node selection strategy and then use the selected seed nodes as cores to attract neighboring nodes into initial communities by incorporating embedding-based similarity. This approach not only addresses the instability of community detection caused by random seed nodes but also enhances the adaptability and stability of the algorithm in both sparse and dense networks. Building upon this foundation, we develop a boundary node-based local random walk attraction method to further expand community ranges and accurately capture nodes in overlapping regions. This method identifies and incorporates nodes located at community boundaries, reduces the introduction of peripheral “noisy nodes” and avoids the isolation or misclassification of boundary nodes. The overall algorithm balances node structural characteristics with propagation paths, effectively improving the accuracy and robustness of community detection. Experimental results on multiple synthetic and real-world networks demonstrate that SSFBW is applicable to detecting overlapping community structures and outperforms existing mainstream methods in terms of stability and feasibility.

A boy wants to make friends with a pretty girl. He feels that he may get rejected if he invites her directly. In this situation, what he could do is to influence the girl's friends. Similar situations may occur in social activities. Based on this background, we formulate a new optimization problem, the Target Influence Maximization (TIM) problem and show that this problem can be solved in polynomial-time in networks with no directed cycles. Motivated by this, we study a special strategy to construct solutions for TIM, i.e., The Target Influence Maximization through Sub graph without Directed Cycle (TIMSDC). Two polynomial-time approximation algorithms are designed for TIMSDC. Through extensive experiments on real-world data sets, we demonstrate that our algorithms work efficiently and outperform existing methods.

Describing the behavior of information dissemination in online social networks is propitious to understanding the transmission process of real users in the online social networking site. In this paper, we proposed a cellular automaton model to deal with the propagation characteristics of online social networks rumors spread. Experimental simulation was carried out under periodic boundary constraints in the process of rumor spread. The result showed that the cellular automaton model is indeed able to characterize the propagation behavior on online social networks. We also prescribed the immunization strategy to suppress the rumor spreading.

The interaction between individuals are usually modeled as weighted edges in a social network. This information, however, is often unavailable in practice. On the other hand, information diffusion process upon the underlying network is observable. Hence sophisticated algorithm is needed to infer the edge set and edge weights from observed cascade set. To deal with this problem, we derive the likelihood of a given network generating a cascade set. With this likelihood, we design a distributed algorithm named Net Win that first calculates the optimal edge weights by maximizing likelihood and then sparsifies the result of optimization by a novel post-processing algorithm. In experimental results, Net Win infers various networks with high accuracy and outperforms other state-of-the-art algorithms in almost all cases.