Link prediction for new users in Social Networks

Abstract

International audienceLink prediction for new users who have not created any link is a fundamental problem in Online Social Networks (OSNs). It can be used to recommend friends for new users to start building their social networks. The existing studies use crossplatform approaches to predict a new user's links on a certain OSN by porting his existing links from other OSNs. However, it cannot work when OSNs are not willing to share their data or users do not want to connect different OSN accounts. In this paper, we use a single-platform approach to carry out the link prediction. We explore the users' profile attributes (e.g., workplace, high school and hometown) which can be easily obtained during the new users' sign up procedure. Based on the limited available information from the new user, along with the attributes and links from existing users, we extract three types of social features: basic feature, derived feature and latent relation feature. We propose a link prediction model using these social features based on Support Vector Machines. Eventually, we rely on a large Facebook data set consisting of 479; 000 users to evaluate our proposed model. The result reveals that our model outperforms the baselines by achieving the AUC value of 0:83; it also demonstrates that each of the proposed social features contribute significantly to the prediction mode

Missing Link Prediction Using Non-Overlapped Features and Multiple Sources of Social Networks

“…Facebook latent (FL) is the most effective feature from the linking prediction research proposed by Han et al [12]. The principle of this feature is to consider the possibility of a relationship between two nodes by considering the profile features from the first-order neighbors compared with the profile features of the other considered nodes.…”

“…where r is the number of similar features (the number of latent links), q is the number of different features (the number of disconnections), β is an exponential regulator, and α is a regulator for punishing value [24]. According to Han et al [12], β equals 0.05 and α equals 0.05. (2) When considering the qualitative features of nodes on Twitter, the features of each node on Twitter consist of data that each node tweets, which contains hashtag data and the mentioned users.…”

“…The machine-learning approach applies the existing relationship between nodes and links from a dataset to create a classification model. There are several works [11][12][13][14][15][16][17][18][19][20][21] in this approach but with different machine-learning techniques and SN sources. To increase an efficiency of the link prediction, some works exploit user profile features such as age, gender, and current city along with graph-based relationships as training features.…”

“…RPF explains the probability that two nodes will know each other when many responses are between two nodes. This can be calculated from Equation (12),…”

Extending adamic adar for cold-start problem in link prediction based on network metrics

“…Furthermore, Han et al [28] examined the users' attributes and proposed new users prediction using users' social features based on the Support Vector Machine. Wang et al [29] also presented a possible connection between cold-start users and existing users based on topological information extraction.…”