Package es.uam.eps.ir.relison.links.recommendation.algorithms.standalone.foaf
Classes and definitions of algorithms based on friends of friends link prediction approaches.
-
Class Summary Class Description AdamicAdar<U> Recommender that uses the Adamic-Adar coefficient of the neighbours.Cosine<U> Recommender using the cosine similarity to produce recommendations.Dist2Popularity<U> Recommender that sorts users at distance 2 by popularity.HubDepressedIndex<U> Recommender that uses the hub depressed index of the neighbors: given the number of common neighbors between two users, the recommendation score is divided by the size of either the target user or the candidate user: the user with a larger number of them.HubPromotedIndex<U> Recommender that uses the hub depressed index of the neighbors: given the number of common neighbors between two users, the recommendation score is divided by the size of either the target user or the candidate user: the user with a smaller number of them.Jaccard<U> Recommended based on the Jaccard similarity.LocalLHNIndex<U> Recommender that uses the local Leicht-Holme-Newman index.MostCommonNeighbors<U> Recommended that sorts candidate users according to the number of neighbors in common with the target one.NonReciprocalPreferentialAttachment<U> Non Reciprocal Preferential Attachment recommender.PreferentialAttachment<U> Recommender based on the preferential attachment phenomena.ReciprocalLinks<U> Recommends reciprocal links.ResourceAllocation<U> Recommender that uses the resource allocation principle to recommend.Sorensen<U> Recommender based on Sorensen similarity.TotalNeighbors<U> Recommends people according to the total number of neighbors between the two users.