Check goodness of fit for 'nomad_model' class

This function takes a nomad_model object and calculates goodness of fit metrics for the underlying mobility::mobility(). If the Deviance Information Criterin (DIC) was calculated in the supplied model object, it is included in output. When plots = TRUE, two plots are shown containing the posterior distribution of trip counts compared to observed data and a Normal Q-Q plot showing the quantiles of model residuals against those expected from a Normal distribution.

check(object, plots, ...)

Arguments

object

a nomad_model() object containing the fitted mobility

plots

logical indicating whether to plot the Posterior Predictive Check and Normal Q-Q Plot (default = TRUE)

...

further arguments passed to or from other methods

Value

a list of goodness of fit measures

Details

Goodness of fit metrics include:

DIC

Deviance Information Criterion

RMSE

Root Mean Squared Error

MAPE

Mean Absolute Percent Error

R2

R-squared

See also

Other model: compare(), fit_jags(), fit_prob_travel(), mobility(), predict(), residuals(), summary()

Author

John Giles

Examples

# Get nomad_model object
nmd_model <- nomad::model_db$zmb_cdr_2020_mod_dd_exp

# Check model fit
nomad::check(nmd_model)

#> $DIC
#> [1] 76328739
#> 
#> $RMSE
#> [1] 172445.4
#> 
#> $MAPE
#> [1] 23473.45
#> 
#> $R2
#> [1] 0.7952939
#> 

# Get nomad_model object without underlying data
nmd_model <- nomad::model_db$zmb_fb_2020_mod_grav_exp

# Model check statistics are still available as these are
# saved when model data is removed from object
nomad::check(nmd_model)

#> $DIC
#> [1] 12442470
#> 
#> $RMSE
#> [1] 447193.4
#> 
#> $MAPE
#> [1] 55.19494
#> 
#> $R2
#> [1] 0.3721586
#>