Xiang et al., 2010,Modeling Relationship Strength in Online Social Networks
Contents
Online version
An online version of this paper is available at the [ACM digital library].
Summary
This paper investigates unsupervised models for Determining Social Network Attributes, more specifically, link strength in social network. Previous work focusing on friendship relations mostly assumes binary relation (connected or not connected). However, the authors argue that real-life network is a more complicated environment, where acquaintances and best-friends relations are mixed together. They develop an unsupervised model to estimate the strenght of these relations by using features such as bi-directional communication as well as user similarity. Their approach is evaluated on Facebook, and shows an improved classification accuracy.
Key Contributions
Unsupervised model for predicting relationship strength.
Background
This work is based on the principle of homophily, which states that two persons who have similar characteristics tend to tie to each other more strongly than two persons with no similarity. This relationship strength is assumed to impact directly the frequency of online communications, such as emails and direct messages in Facebook.
Models
Latent variable model
The task of predicting edge sign (positive or negative) is accomplished by a logistic regression model that uses the following features:
- signed degree of a node (i.e., how many positive edges it has)
- "type" of triads formed by (u,v and w) in a way that the node w has an edge to both u and v. For example, a negative/negative relations exists if w foe with both u and v (thus u and v might be friends, as explained below). There are 16 types of such triads.
Experiments and Evaluation
The model is compared to a random approach, where positive and negative are randomly selected. For this approach to be valid, the authors sample from the SlashDot dataset to obtain 50% of positive and negative. The effects of the different features on the logistic regression are shown in the following results graph:
Source: the original paper
In this graph, A corresponds to predicting Epinions "distrust" relationships, B corresponds to SlashDot "foes" relationship and C corresponds to a user voting down another user for adminship in Wikipedia. The results indicate that it is possible to predicting the polarity of a link with high accuracy compared to a random baseline. This results seems to generalize to multiple dataset.
