LCD: IDENTIFYING INFLUENTIAL NODES IN COMPLEX NETWORKS WITH LAYERED CLUSTERING AND DEGREE

Identifying influential nodes in networks is a key challenge in understanding how information spreads. While numerous algorithms have been proposed in the literature, many struggle with either limited spreading efficiency or high computational complexity. To address this challenge, we present Layered Clustering Degree (LCD), an effective method for identifying a set of well-distributed spreaders with high spreading ability, while maintaining a computational complexity of making it highly suitable for large-scale networks where both efficiency and computational complexity are essential. The LCD approach operates in three main steps: (1) Layering, which organizes nodes hierarchically based on their shortest distances from a designated starting node; (2) Clustering, which groups nodes within each layer into connected substructures to capture local connectivity patterns; and (3) Degree computation and ranking, where node degrees are computed within the entire network (globally) but ranked iteratively across clusters (locally) to ensure maximum coverage and minimal overlap among selected spreaders. The significance of layered clustering in LCD method, is iteratively distributing spreaders across clusters to avoid over-representation of high-degree nodes from a single region of the network. Experimental results using the SIR model on nine real-world networks show that LCD outperforms several popular methods, including VoteRank, K-shell, VoteRank++, ClusterRank, H-Index, EnRenew, and DegreeRank, in terms of spreading rate and final affected scale