@article{fdi:010060673,
  title = {{H}idden {M}arkov {M}odels : the best models for forager movements ?},
  author = {{J}oo, {R}. and {B}ertrand, {S}ophie and {T}am, {J}. and {F}ablet, {R}.},
  editor = {},
  language = {{ENG}},
  abstract = {{O}ne major challenge in the emerging field of movement ecology is the inference of behavioural modes from movement patterns. {T}his has been mainly addressed through {H}idden {M}arkov models ({HMM}s). {W}e propose here to evaluate two sets of alternative and state-of-the-art modelling approaches. {F}irst, we consider hidden semi-{M}arkov models ({HSMM}s). {T}hey may better represent the behavioural dynamics of foragers since they explicitly model the duration of the behavioural modes. {S}econd, we consider discriminative models which state the inference of behavioural modes as a classification issue, and may take better advantage of multivariate and non linear combinations of movement pattern descriptors. {F}or this work, we use a dataset of >200 trips from human foragers, {P}eruvian fishermen targeting anchovy. {T}heir movements were recorded through a {V}essel {M}onitoring {S}ystem (similar to 1 record per hour), while their behavioural modes (fishing, searching and cruising) were reported by on-board observers. {W}e compare the efficiency of hidden {M}arkov, hidden semi-{M}arkov, and three discriminative models (random forests, artificial neural networks and support vector machines) for inferring the fishermen behavioural modes, using a cross-validation procedure. {HSMM}s show the highest accuracy (80%), significantly outperforming {HMM}s and discriminative models. {S}imulations show that data with higher temporal resolution, {HSMM}s reach nearly 100% of accuracy. {O}ur results demonstrate to what extent the sequential nature of movement is critical for accurately inferring behavioural modes from a trajectory and we strongly recommend the use of {HSMM}s for such purpose. {I}n addition, this work opens perspectives on the use of hybrid {HSMM}-discriminative models, where a discriminative setting for the observation process of {HSMM}s could greatly improve inference performance.},
  keywords = {},
  booktitle = {},
  journal = {{PL}o{S} {O}ne},
  volume = {8},
  numero = {8},
  pages = {71246 [13 ]},
  ISSN = {1932-6203},
  year = {2013},
  DOI = {10.1371/journal.pone.0071246},
  URL = {https://www.documentation.ird.fr/hor/fdi:010060673},
}