LinkPred: a high performance library for link prediction in complex networks

Abstract

The problem of determining the likelihood of the existence of a link between two nodes in a network is called link prediction. This is made possible thanks to the existence of a topological structure in most real-life networks. In other words, the topologies of networked systems such as the World Wide Web, the Internet, metabolic networks, and human society are far from random, which implies that partial observations of these networks can be used to infer information about undiscovered interactions. Significant research efforts have been invested into the development of link prediction algorithms, and some researchers have made the implementation of their methods available to the research community. These implementations, however, are often written in different languages and use different modalities of interaction with the user, which hinders their effective use. This paper introduces LinkPred, a high-performance parallel and distributed link prediction library that includes the implementation of the major link prediction algorithms available in the literature. The library can handle networks with up to millions of nodes and edges and offers a unified interface that facilitates the use and comparison of link prediction algorithms by researchers as well as practitioners.

RELISON: A Framework for Link Recommendation in Social Networks

“…More closely related to our framework, several representative link prediction libraries have been released: LPMade [67] provides some classical unsupervised and supervised link prediction algorithms, whereas LinkPred [58] includes some classical approaches along with more recent methods based on graph embedding algorithms as node2vec [42]. These link prediction libraries support the classification view of the problem -they do not supply code for running and evaluating predictions in recommendation mode.…”

“…Our framework thus extends currently available recommender system frameworks with specific methods for recommending people in networks. Conversely, currently available software addressing the link prediction problem [54,58,67] do not support the specific formulation and methodology to apply prediction as a recommendation task.…”

A Heterogeneous Graph Model for Repurposing Drugs against Echinococcosis

“…ROC curves are also used to evaluate how true or false a predictor is. We made our algorithms in the Python programming language and developed them using the LinkPred [13] library. As can be seen from the ROC curve shown in Figure 6, for the three selected predictors, the Common Neighbor predictor predicts slightly more efficiently by considering the common neighboring vertices.…”