The accelerating growth of online tagging systems

Abstract

Research on the growth of online tagging systems not only is interesting in its own right, but also yields insights for website management and semantic web analysis. Traditional models that describing the growth of online systems can be divided between linear and nonlinear versions. Linear models, including the BA model (Brabasi and Albert, 1999), assume that the average activity of users is a constant independent of population. Hence the total activity is a linear function of population. On the contrary, nonlinear models suggest that the average activity is affected by the size of the population and the total activity is a nonlinear function of population. In the current study, supporting evidences for the nonlinear growth assumption are obtained from data on Internet users' tagging behavior. A power law relationship between the number of new tags (F) and the population (P), which can be expressed as F ~ P ^ gamma (gamma > 1), is found. I call this pattern accelerating growth and find it relates the to time-invariant heterogeneity in individual activities. I also show how a greater heterogeneity leads to a faster growth.