Abstract by Joseph Leung
Utilizing Group Affinity to Predict Community Formation in Social Networks
The way groups form over time is of great interest to many fields, particularly in the study of social networks. A challenge in this area is predicting which individuals will at some future time belong to a specific community based on their previous social interactions. For a number of community detection methods, we derive a corresponding affinity score for each individual in a social network. We use this affinity along with a number of Machine Learning Algorithms (MLA) to predict communities at a later moment in time. We show that certain affinity scores have the capability of correctly predicting group formation in real-world social networks effectively.