Table 2 Number of estimated parameters (K), AIC, and ΔAIC (compared with best model) for 3-state HMMs with different covariate dependences during general phase for residents and non-residents, based on 1-h interval dataset. The effect of the time of day is cyclical over 24 h. For both data sets, the model with all three covariates was selected by the AIC
From: Understanding decision making in a food-caching predator using hidden Markov models
K | AIC | ΔAIC | Model weight | |
|---|---|---|---|---|
Resident individuals | ||||
Time of day + Temperature + Hunger | 47 | 6257.7 | 0 | 1 |
Time of day + Temperature | 41 | 6285.2 | 27.5 | 0 |
Time of day | 35 | 6312.5 | 54.8 | 0 |
No covariate | 23 | 6370.1 | 112.3 | 0 |
Non-resident individuals | ||||
Time of day + Temperature + Hunger | 44 | 5849.7 | 0 | 1 |
Time of day + Temperature | 38 | 5900.7 | 50.9 | 0 |
Time of day | 32 | 5982.3 | 81.6 | 0 |
No covariate | 20 | 6176.0 | 193.7 | 0 |