Time Score: A New Feature for Link Prediction in Social Networks

Abstract

Lankeshwara MUNASINGHE†a) , Nonmember and Ryutaro ICHISE †b) , Member SUMMARY Link prediction in social networks, such as friendship networks and coauthorship networks, has recently attracted a great deal of attention. There have been numerous attempts to address the problem of link prediction through diverse approaches. In the present paper, we focus on the temporal behavior of the link strength, particularly the relationship between the time stamps of interactions or links and the temporal behavior of link strength and how link strength affects future link evolution. Most previous studies have not sufficiently discussed either the impact of time stamps of the interactions or time stamps of the links on link evolution. The gap between the current time and the time stamps of the interactions or links is also important to link evolution. In the present paper, we introduce a new time-aware feature, referred to as time score, that captures the important aspects of time stamps of interactions and the temporality of the link strengths. We also analyze the effectiveness of time score with different parameter settings for different network data sets. The results of the analysis revealed that the time score was sensitive to different networks and different time measures. We applied time score to two social network data sets, namely, Facebook friendship network data set and a coauthorship network data set. The results revealed a significant improvement in predicting future links.

Exploring Supervised Methods for Temporal Link Prediction in Heterogeneous Social Networks

“…There are already works which incorporate time information into the calculation of similarity scores. For example, Munasinghe and Ichise suggest a time-score which combines the time and weight of links with common neighbors [12]. Their approach is based on two concepts: the strength of a link decreases with time, and the common neighbors are more effective if the interaction with them happens in a closer proximity of time.…”

“…Based on the obtained time series for a specific neighborhood based score, they predict the next value of the series and use the predicted value in the link prediction model. However, many works on temporal link prediction consider only homogeneous networks [5,12].…”

“…We introduce three modifications to the original weighting scheme of triads. The first modification is the introduction of time awareness which we achieve by using the time score [12]. The second modification is motivated by the work in the area of graph pattern mining where the stress is put on the fact that simple counts are not always a proper measure to estimate the frequency of patterns in the network [4].…”

Link Prediction in Social Networks Using Information Flow via Active Links

“…The recent research [12] has introduced a new feature which captures the impact of information flow via active links for link evolution in social networks. However, it is limited to common neighbors.…”

“…We used two different combinations of features in the proposed machine learning approach for link prediction. The two sets of features includes a set of features used in [12] with PropFlow score computed by previous PropFlow algorithm [10] and T Flow score computed by T Flow algorithm introduced in this paper. One set was used as the PropFlow combination which includes the PropFlow score and used as the base line combination.…”