This is LinkPred, a high performance library for link prediction in complex networks.
LinkPred provides the following functionalities:
Basic data structures to efficiently store and access network data.
Basic graph algorithms such graph traversal, shortest path algorithms, and graph embedding methods.
Implementation of several topological similarity index predictors, for example: common neighbors, Adamic-Adard index and Jackard index among other predictors (a full list is available in the library documentation).
Implementation of several state-of-the-art global link predictors (a full list is available in the library documentation).
Implementation of several link prediction algorithms based on graph embedding techniques.
Test data generation from ground truth networks.
Performance evaluation functionalities.