Community Detection Using Fireworks Optimization Algorithm

Since community detection is an important tool for understanding the complex structure of social networks, an improved fireworks algorithm is proposed in this paper. The algorithm generates the initial population with the Affinity Propagation approach to have high initialization quality. The algorithm optimizes the modularity density as objective function by calculating the amplitude, the number of sparks and exploring the sparks. One firework is mutated twice, randomly and according to the label of its neighbors. Experiments on both real and synthetic networks show that the proposed algorithm achieves more accurate results in terms of modularity and normalized mutual information.