Cross-validation procedure

Cross-validation for estimation of all AMMI-family models

cv_ammif provides a complete cross-validation of replicate-based data using AMMI-family models. By default, the first validation is carried out considering the AMMIF (all possible axis used). Considering this model, the original dataset is split up into two datasets: training set and validation set. The 'training' set has all combinations (genotype x environment) with N-1 replications. The 'validation' set has the remaining replication. The splitting of the dataset into modeling and validation sets depends on the design informed. For Completely Randomized Block Design (default), and alpha-lattice design (declaring block arguments), complete replicates are selected within environments. The remained replicate serves as validation data. If design = 'RCD' is informed, completely randomly samples are made for each genotype-by-environment combination (Olivoto et al. 2019). The estimated values for each member of the AMMI-family model are compared with the 'validation' data. The Root Mean Square Prediction Difference (RMSPD) is computed. At the end of boots, a list is returned.

IMPORTANT: If the data set is unbalanced (i.e., any genotype missing in any environment) the function will return an error. An error is also observed if any combination of genotype-environment has a different number of replications than observed in the trial.

Usage

cv_ammif(
  .data,
  env,
  gen,
  rep,
  resp,
  nboot = 200,
  block,
  design = "RCBD",
  verbose = TRUE
)

Arguments

.data: The dataset containing the columns related to Environments, Genotypes, replication/block and response variable(s).
env: The name of the column that contains the levels of the environments.
gen: The name of the column that contains the levels of the genotypes.
rep: The name of the column that contains the levels of the replications/blocks. AT LEAST THREE REPLICATES ARE REQUIRED TO PERFORM THE CROSS-VALIDATION.
resp: The response variable.
nboot: The number of resamples to be used in the cross-validation. Defaults to 200.
block: Defaults to NULL. In this case, a randomized complete block design is considered. If block is informed, then a resolvable alpha-lattice design (Patterson and Williams, 1976) is employed. All effects, except the error, are assumed to be fixed.
design: The experimental design used in each environment. Defaults to RCBD (Randomized complete Block Design). For Completely Randomized Designs inform design = 'CRD'.
verbose: A logical argument to define if a progress bar is shown. Default is TRUE.

Value

An object of class cv_ammif with the following items:

RMSPD: A vector with nboot-estimates of the Root Mean Squared Prediction Difference between predicted and validating data.
RMSPDmean: The mean of RMSPDmean estimates.
Estimated: A data frame that contain the values (predicted, observed, validation) of the last loop.
Modeling: The dataset used as modeling data in the last loop
Testing: The dataset used as testing data in the last loop.

References

Patterson, H.D., and E.R. Williams. 1976. A new class of resolvable incomplete block designs. Biometrika 63:83-92.

Author

Tiago Olivoto tiagoolivoto@gmail.com

Examples


# \donttest{
library(metan)
model <- cv_ammif(data_ge2,
                  env = ENV,
                  gen = GEN,
                  rep = REP,
                  resp = EH,
                  nboot = 5)
#> 1 of 5 sets using AMMIF |==                                      | 5% 00:00:00 
2 of 5 sets using AMMIF |====                                    | 10% 00:00:00 
3 of 5 sets using AMMIF |======                                  | 15% 00:00:00 
4 of 5 sets using AMMIF |========                                | 20% 00:00:00 
5 of 5 sets using AMMIF |==========                              | 25% 00:00:01 
1 of 5 sets using AMMI2 |============                            | 30% 00:00:01 
2 of 5 sets using AMMI2 |==============                          | 35% 00:00:01 
3 of 5 sets using AMMI2 |================                        | 40% 00:00:01 
4 of 5 sets using AMMI2 |==================                      | 45% 00:00:02 
5 of 5 sets using AMMI2 |====================                    | 50% 00:00:02 
1 of 5 sets using AMMI1 |======================                  | 55% 00:00:02 
2 of 5 sets using AMMI1 |========================                | 60% 00:00:02 
3 of 5 sets using AMMI1 |==========================              | 65% 00:00:02 
4 of 5 sets using AMMI1 |============================            | 70% 00:00:03 
5 of 5 sets using AMMI1 |==============================          | 75% 00:00:03 
1 of 5 sets using AMMI0 |================================        | 80% 00:00:03 
2 of 5 sets using AMMI0 |==================================      | 85% 00:00:03 
3 of 5 sets using AMMI0 |====================================    | 90% 00:00:03 
4 of 5 sets using AMMI0 |======================================  | 95% 00:00:04 
5 of 5 sets using AMMI0 |========================================| 100% 00:00:04 

plot(model)

# }