Abstract
When analysing a real-world problem, it can be transformed into a complex network, and then its structure is divided to find out the overlapping part and study the hidden laws. In this paper, three overlapping community recognition algorithms are briefly introduced, i.e., overlapping community partitioning based on label propagation (OCPLP), footpad skin optical clearing agent (FSOCA) and the genetic algorithm (GA), and then the three artificial networks generated by the LFR tool are simulated and analysed through matrix laboratory (MATLAB) software. The results show that the increase of network complexity will reduce the recognition performance of the three algorithms. In the three algorithms, GA is relatively less affected, which always has the highest recognition performance. OCPLP is the most affected, which has the worst performance. In terms of the speed of recognition, GA is also the fastest, and OCPLP is the slowest. In summary, GA is more suitable for the search of overlapping communities in the network than that of OCPLP and FSOCA.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call