eval_model: Evaluating Models according to Residuals

In kamapu/cropgrowth: Assessing Results from Crop Growth Modelling

Description

Evaluation of models will be done comparing observed with simulated data. Simulated data should be of a higher resolution than observed, while the time of sampling may not match exactly the simulation steps. Additionally, multiple observations (repetitions) may be compared with single simulations.

This function matches predictions and observations according to the variable 'match' within the samples belonging to the same group (i.e. treatment).

No functions are included to contrast predictions with observations, thus they have to be defined by the user. Those functions have to include two arguments, namely 'x' for predictions and 'y' for observations.

Usage

eval_model(pred, obs, vars, match, group, FUN = list(), na.rm = FALSE, ...)

Arguments

pred, obs	Data frames containing predicted and observed variables, respectively.
vars	Character vector with the names of variables to be evaluated.
match, group	Character values or vectors indicating the names of variables used to match observations with predictions and to group treatments (see also match_observations).
FUN	List with functions used for the evaluation.
na.rm	Logical value, whether NAs in predictions and observations have to be skipped or not.
...	Further arguments passed to approx (see also match_observations).

Value

A data frame.

References

Janssen PHM, Heuberger PSC (1995). Calibration of process-oriented models. Ecological Modelling 83: 55–66. https://doi.org/10.1016/0304-3800(95)00084-9

Bennett ND et al. (2013). Characterising performance of environmental models. Environmental Modelling and Software 40: 1–20. http://dx.doi.org/10.1016/j.envsoft.2012.09.011

Examples

## No example at the moment

kamapu/cropgrowth documentation built on Aug. 22, 2021, 8:55 a.m.