Skip to main content

Table 5 Variation in entropy explained by each of the proposed regression models (adjusted \(R^{2}\)) for both datasets, in the next-place prediction task. The R model is the model that uses regularity, and the RD model is the one where diversity of trajectories is also used, along with regularity. We also include results for the RDN model, which in addition to regularity and diversity also uses the size n of ones routine, and the RDNB model, which adds information about the baseline entropy of one’s routine

From: On estimating the predictability of human mobility: the role of routine

 

GPS dataset

CDR dataset

 

Adjusted \(R^{2}\)

Adjusted \(R^{2}\)

R Model

0.672

0.735

RN Model

0.723

0.801

RND Model

0.739

0.852

RNDB Model

0.750

0.855