Computationally efficient link prediction in a variety of social networks

Abstract

Online social networking sites have become increasingly popular over the last few years. As a result, new interdisciplinary research directions have emerged in which social network analysis methods are applied to networks containing hundreds of millions of users. Unfortunately, links between individuals may be missing either due to an imperfect acquirement process or because they are not yet reflected in the online network (i.e., friends in the real world did not form a virtual connection). The primary bottleneck in link prediction techniques is extracting the structural features required for classifying links. In this article, 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 predicament of classifying links between individuals with at least one common friend. We also present a method for calculating the amount of data needed in order to build more accurate classifiers. The new Friends measure and Same community features we developed are shown to be good predictors for missing links. An evaluation experiment was performed on ten large social networks datasets: Academia.edu, DBLP, Facebook, Flickr, Flixster, Google+, Gowalla, TheMarker, Twitter, and YouTube. 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 online social networks.

A Novel Strategy for Link Prediction in Social Networks

“…Nodes 2 and 5 has got five and four neighbors respectively amongst which 2 neighbors are common to both the nodes. So, Γ(2) = (1,3,4,7,8) and Γ(5) = (4,6,7,9).…”

“…In the field of electronic commerce, it can be used to create the recommender systems; and in the security field, it can help to find the hidden terrorist and criminal gangs. Therefore, in recent years a lot of algorithms have been proposed to solve the problem of link prediction [1,4,5,6,7,12,13,14,15,16,19,20,22].…”

ConSent: Context-based sentiment analysis

“…As our key and context terms can be represented using a graph (as shown in Fig. 2), our aim is to use measures from graph theory such as betweenness and centrality [95][96][97] to enrich our feature set. Another interesting research prospect would be to adapt our method to multi-class problems (where there is a need to classify more than two item types simultaneously)-both in the field of sentiment analysis and in other classification tasks.…”

Detecting partnership in location-based and online social networks

“…For this task, we applied four well-known algorithms on the previously mentioned balanced dataset with 1500 partner pairs and 1500 non-partner pairs which were (see also Fire et al 2014): J.48, Logistic Regression, SVM (radial kernel) and Bagging (with RepTree classifier) using the WEKA machine learning Hall et al 2009 suite using tenfold cross validation. Due to the balanced nature of the dataset comprising both pairs of users who were in a partnership and those who were not (non-partners), the baseline for the classification task was 0.5 AUC when guessing at random.…”