Fig 6. Dynamics of daily tweet volume for each party represented by data and by the GBM model.

In the GBM model, the expected volume V(t) at time t is given by $V_p(t)=V_p(0)exp((\mu -\frac{{\sigma }^2}{2})t)$. In the GBM+σ model, $V_p(t)=V_p(0)exp((\mu -\frac{{\sigma }^2}{2})t+\sigma )$ while $V_p(t)=V_p(0)exp((\mu -\frac{{\sigma }^2}{2})t-\sigma )$ in the GBM−σ model. The values of parameters μ, σ, and V(0) are given in Table 3.