Influence maximization in blockchain social networks: a heterogeneous LDAG approach

In the Web 3.0 epoch, the evolution of blockchain technology has precipitated the advent and expansion of Blockchain Online Social Networks (BOSNs). These decentralized networks vow to augment data privacy, security, and user autonomy, concurrently presenting analogous challenges in technology. In this paper, we introduce a Heterogeneous Local Directed Acyclic Graph (HLDAG) algorithm, designed to tackle the challenge of influence maximization in BOSNs. Leveraging the complex structure of heterogeneous information networks, HLDAG constructs local directed acyclic graphs for target nodes to efficiently measure and maximize influence spread. Based on the linear threshold model, we optimized the influence model and threshold model using BOSN social data. HLDAG assigns weights to paths and nodes based on their semantic significance, effectively guiding the process of identifying key influencers in the network. The algorithm’s performance was evaluated using data from Steemit, a decentralized platform representative of BOSNs. Various experiments are designed to evaluate the performance of the proposed model and algorithm. Experimental results demonstrate HLDAG’s superior efficiency and accuracy in influence spread estimation compared to existing algorithms, highlighting its potential in enhancing data security governance in Web 3.0.