A Link Prediction Similarity Index Based on Enhanced Local Path Method

The index of link prediction based on local path considers the contribution of third order neighbors on the basis of the common neighbor index, which has strong randomness and ignores the influence of the particularity of network topology on particle transition probability. In order to resolve this problem, this paper proposes a enhanced local path (ELP) link prediction similarity index. The algorithm considers the contribution of the third order path should have the relationship with the degree of the nodes at both ends of the path. By calculating the similarity of node vectors, it redefines a similarity index based on enhanced local path. We apply it to four different types of networks. Finally, the index proposed in this paper is compared with five classical similarity indexes. The results show that the prediction algorithm based on enhanced local path index has higher accuracy than other indexes.