Fig. 6
From: Modelling sequences and temporal networks with dynamic community structures
Markov model fit for a concatenated text. ‘The texts are ‘War and peace’ by Leo Tolstoy and ‘À la recherche du temps perdu’, by Marcel Proust. Edge endpoints in red and blue correspond to token and memories, respectively, that involve letters exclusive to French. a Version of the model with n = 3 where no distinction is made between the same token in the different novels, yielding a description length \( - {\rm{lo}}{{\rm{g}}_{\rm{2}}}\,P\left( {\left\{ {{x_t}} \right\},b} \right) \) = 7,450,322. b Version with n = 3 where each token is annotated with its novel, yielding a description length \( - {\rm{lo}}{{\rm{g}}_{\rm{2}}}\,P\left( {\left\{ {{x_t}} \right\},b} \right) = 7,146,465\)
