Student similarity network clustering -- Does the time it takes for an answer to be selected follow a power-law?
This article uses a dataset of answers to questions to generate student similarity networks. Two similarity functions to determine the weights between each pair of students are used, one that assumes a power-law distribution of answers response times, and one that does not. The resulting networks are then clustered using different community finding algorithms and their resulting modularity is compared.
physics.soc-ph↗