Community-based Influence Maximization Using Network Embedding in Dynamic Heterogeneous Social Networks

Influence maximization (IM) is a very important issue in social network diffusion analysis. The topology of real large-scale, dynamic, and heterogeneous. heterogeneity, continuous expansion evolution pose challenge to find influential users. Existing IM algorithms usually assume that networks are static or dynamic but homogeneous simplify the complexity problem. We propose community-based influence algorithm using embedding heterogeneous networks. use DyHATR obtain propagation feature vectors nodes, execute k -means cluster transform original into coarse granularity (CGN). On CGN, we three-hop independent cascade model construct objective function design greedy heuristics solve problem with \((1-\frac{1}{e})-\) approximation guarantee community structure quickly identify seed users estimate their value. Experimental results on demonstrated compared existing algorithms, our proposed had better comprehensive performance respect value, more less execution time memory consumption, scalability.