Link Prediction

Link Prediction in Highly Fractional Data Sets

by Michael Fire

co-authored with "Rami Puzis and Yuval Elovici, draft version

Extremist organizations all over the world increasingly use online social networks as a communication media for... more Extremist organizations all over the world increasingly use online social networks as a communication media for recruitment and planning. As such, online social networks are also a source of information utilized by intelligence and counter
terror organizations investigating the relationships between suspected individuals. Unfortunately, the data mined from open sources is usually far from being complete due to the efforts of suspected and known terrorists to hide their relationships. One
of the methods used to uncover missing information in social networks is referred to as link prediction. We use link prediction methods solely based on network struc-ture analysis to infer hidden relationships among individuals and investigate their
effectiveness in fractional datasets. Experiments performed on a number of closed communities extracted from organizational and public social networks show that structural link prediction retains its effectiveness even when large parts of the origi-nal social network are hidden.

Links Reconstruction Attack Using Link Prediction Algorithms to Compromise Social Networks Privacy

by Michael Fire

Co-authored with Gilad Katz, Lior Rokach and Yuval Elovici

The explosion in the use of social networks has led to the creation of new kinds of security and privacy threats. Many... more The explosion in the use of social networks has led to the creation of new kinds of security and privacy threats. Many users are unaware of the risks involved with exposing personal information, a fact that makes social networks a “bonanza” for spammers and identity thieves. In addition, it has already been proven that even the concealment of all personal data might not be enough to provide protection, as one’s personal information can be inferred by analyzing one’s connections to other users. In this paper we present the “link reconstruction attack”, a method capable of inferring one’s connections to others with high accuracy. We show that the concealment of ones links is ineffective if not done by others in the network and present an analysis of the performance of various machine learning algorithms for link prediction inside communities.

The link prediction problem in bipartite networks

by Jérôme Kunegis

Link Prediction in Social Networks using Computationally Efficient Topological Features

by Michael Fire

Contians refrence to the following social network datasets: Facebook, Flickr, TheMarker, Academia.edu and YouTube

Online social networking sites have become increas-ingly popular over the last few years. As a result, new... more Online social networking sites have become increas-ingly popular over the last few years. As a result, new interdisci-plinary research directions have emerged in which social network analysis methods are applied to networks containing hundreds millions of users. Unfortunately, links between individuals may be missing either due to imperfect acquirement processes or because they are not yet reflected in the online network (i.e., friends in real-world did not form a virtual connection.) Existing link prediction techniques lack the scalability required for full application on a continuously growing social network. The primary bottleneck in link prediction techniques is ex-tracting structural features required for classifying links. In this paper we propose a set of simple, easy-to-compute structural features, that can be analyzed to identify missing links. We show that by using simple structural features, a machine learning classifier can successfully identify missing links, even when applied to a hard problem of classifying links between individuals
with at least one common friend. A new friends measure that
we developed is shown to be a good predictor for missing
links. An evaluation experiment was performed on five large
Social Networks datasets: Facebook, Flickr, YouTube, Academia and TheMarker. Our methods can provide social network site operators with the capability of helping users to find known, offline contacts and to discover new friends online. They may also be used for exposing hidden links in an online social network.