Eigenvector label propagation algorithm for interactive learning in student groups based on student social network

摘要

Interactive learning, which is based on data mining, is a hot issue and has attracted considerable attention recently. in this paper, we propose an Eigenvector Label Propagation Algorithm (ELPA), which is improved from Label Propagation Algorithm and solves three problems existing in original algorithm. The efficiency is improved greatly because of the introducing of 2-bit eigenvector label, which can reduce the size of exchanging data significantly. We compare the ELPA with GN and BMLPA on two famous benchmarks, and the experimental results show that the groups detected by ELPA are almost identical to the communities discovered by other LPA algorithms.