# Description of the functions related to the model building task

 runModelBuilding Run the model building task. getModelBuildingSettings Get the settings that will be used during the run of model building. getModelBuildingResults Get the results (detailed models) of the model building. stopModelBuilding Stop the model building task.

### Description

Run model building. Call isRunning to check if the building is still running and get information about the current task, call stopModelBuilding to stop the execution.

### Usage

runModelBuilding(wait = TRUE, ...)


### Arguments

wait
(bool) Should R wait for run completion before giving back the hand to the user. Equals TRUE by default.

...
(list<settings>) Settings to initialize the model building algorithm. See getModelBuildingSettings.

### Details

To change the initialization before a run, use getModelBuildingSettings to receive all the settings. See example.

To launch the function in the background, so that functions which do not modify the project (“get” functions for example) remains available, set the input argument “wait” to FALSE.

getModelBuildingSettings getModelBuildingResults stopModelBuilding

## Not run:

runModelBuilding() # sequential run

set = getModelBuildingSettings()

runModelBuilding(settings = set) # sequential run

runModelBuilding(wait = TRUE) # background run

## End(Not run)

Top of the page, Monolix-R functions

## Get model building settings

### Description

Get the settings that will be used during the run of model building.

### Usage

getModelBuildingSettings()

### Value

The list of settings:

• covariates: (vector<string>) covariate names
• parameters: (vector<string>) parameters names
• strategy: (string) strategy to search best model ([cossac], samba, scm)
• criterion: (string) crtierion to search best model ([BIC], LRT)
• relationships: (data.frame<parameters, covariates, locked>) Use to lock relationships between parameters and covariates. By default, all the combinations are possible. This parameter forces the use or not of some combinations. See example where ka must have SEX and V must not have WEIGHT
• threshold$lrt: threshold used by criterion LRT to continue or not to improve the model (first element is for forward and the second one is for the backward method) • threshold$correlation: threshold used by cossac to choose what combinations (parameter- covariate) must be tried as next candidate model (first element is for forward and the second one is for the backward method)
• useLin: (boolean) computes linearization ([TRUE]) or the Importance Sampling (FALSE)
• useSAMBABeforeCossac: (boolean) gives the possibility to launch one SAMBA iteration before the COSSAC strategy (TRUE, [FALSE])

runModelBuilding

## Not run:

set = getModelBuildingSettings()

set$relationships[1,] = c(“ka”, “SEX”, TRUE) set$relationships[2,] = c(“V”, “WEIGHT”, FALSE)

-> set\$relationships

parameters covariates locked

1 ka SEX TRUE

2 V WEIGHT FALSE

runModelBuilding(settings = set)

## End(Not run)

Top of the page, Monolix-R functions

## Get the results of the model building

### Description

Get the results (detailed models) of the model building

### Usage

getModelBuildingResults()

### Value

The results of model building. All the detailed tried models are returned:

• LL: result of -2*Log-Likelihood
• BICc: modified BIC.
• individualModels: (data.frame) individual model for each individual parameter. The columns are the covariates and the elements of the data.frame notes if a covariate is used or not for the current parameter.

• tested: (vector<string>) first element is the individual parameter and the second one is the covariate. This combination notes if the covariate is tested or not with respect to the previous model.
• bestModel (boolean) best model amongst all the tried models according to the chosen criterion.

SAMBA send the error model and covariance model information if there are exist

• errorModels: chosen type for each error model
• covarianceModels: chosen correlations between individual parameters

runModelBuilding

## Not run:

getModelBuildingResults()

##End(Not run)

## Stop the model building

### Description

Stop the model building.

### Usage

stopModelBuilding()

runModelBuilding