A naïve Bayes model based on overlapping groups for link prediction in online social networks

Abstract

Link prediction in online social networks is useful in numerous applications, mainly for recommendation. Recently, different approaches have considered friendship groups information for increasing the link prediction accuracy. Nevertheless, these approaches do not consider the different roles that common neighbors may play in the different overlapping groups that they belong to. In this paper, we propose a new approach that uses overlapping groups structural information for building a naïve Bayes model. From this proposal, we show three different measures derived from the common neighbors. We perform experiments for both unsupervised and supervised link prediction strategies considering the link imbalance problem. We compare sixteen measures in four well-known online social networks: Flickr, LiveJournal, Orkut and Youtube. Results show that our proposals help to improve the link prediction accuracy.

Exploiting social and mobility patterns for friendship prediction in location-based social networks

“…By using the supervised link prediction it is possible to use different validation techniques, such as k-fold cross-validation. Also, we can use the traditional evaluation measures, such as accuracy, precision, recall, F-measure, AUC, and others, to compare the classifier performance [10], [12].…”

“…In order to observe the impact of the combination of different link prediction methods to make friendship prediction under a supervised context, we considered that the performance of each classifier obtained in each dataset is due to the contribution of the features that constitutes such dataset [11], [12]. Thus, from Table II we observe that in most cases the best AUC is obtained by VTotal, but in some cases VSocial-Locations and VSocial-Proposals achieve the best performance.…”

“…Although there are methods that efficiently combine neighborhood and social groups patterns [11], [12], to the best of our knowledge there is no method combining patterns on both mobility and other kind of user behaviors.…”

Mining user behavior in location-based social networks

“…To the best of our knowledge, we have conducted the first research considering social group as information source in link prediction. Part of this contribution has been previously published in two international conferences VALVERDE-REBAZA et al, 2015).…”

Propagação em grafos bipartidos para extração de tópicos em fluxo de documentos textuais

“…Aprendizado em grafos: Grafos são simplesmente vértices e arestas, porém é uma representação robusta, na qual as relações criadas entre os objetos representados são informações úteis para a tarefa de classificação. Aproveitando isso, foram investigados algoritmos de aprendizado em grafos e desenvolvidos os seguintes trabalhos em colaboração: VALVERDE-REBAZA et al, 2015).…”