Link Prediction for Egocentrically Sampled Networks

Link prediction in networks is typically accomplished by estimating or ranking the probabilities of edges for all pairs nodes. In practice, especially social networks, data are often collected egocentric sampling, which means selecting a subset nodes and recording their edges. This sampling mechanism requires different tools than typical assumption links missing at random. We propose new computationally efficient link algorithm egocentrically sampled underlying probability matrix its row space. empirically evaluate method on several synthetic real-world show that it provides accurate predictions network links. Supplemental materials including code experiments available online.