Adaptive Fusion Link Prediction Algorithm Based on Non-equilibrium Resource Allocation Index

Link prediction could predict the missing links in the networks and the new links in the future network. The traditional link prediction algorithm uses the similarity index to measure different networks. However, the structural features of networks are different, so link prediction using a single index for all networks cannot achieve satisfactory results. This paper uses the heterogeneity to improve the RA index, and applies the adaptive fusion algorithm to the LP index and the improved RA index. The algorithm is verified in six real networks. According to the experimental, it is found that the ANL index is better than many similarity indices, and the algorithm has good prediction accuracy.