Concentration models settings

Calculation settings

Table 39 Calculation settings for module Concentration models.
Name Type Description
Concentration model tier
ConcentrationModelChoice Custom model, or set according to EFSA Guidance 2012. Note: you may need to set the tier separately in sub-modules.
Default concentration model
ConcentrationModelType The concentration model type that will be used as default for all food/substance combinations. If this model type cannot be fitted, e.g., due to a lack of data, a simpler model will be chosen automatically as a fall-back.
Include MRL fallback model

Boolean

Use the MRL as fallback model in case the occurrence data is insufficient for other concentration modelling options.
Restrict LOR imputation to authorised uses

Boolean

Specifies whether imputation of factor x LOR should be limited to authorised uses only.
Censored values replacement
NonDetectsHandlingMethod How to replace censored values (when not co-modelled, as in censored models).
Factor f (f x LOR or f x LOD or LOD + f x (LOQ - LOD)

Numeric

Replace censored values by Limit of reporting (LOR), Non-detects (LOD) or Non-quantifications (LOQ) times this factor. Constant (f), e.g. 0.5.
MRL Factor (f x MRL)

Numeric

Use f x MRL as concentration estimate of the MRL models.
Sample based

Boolean

Include co-occurrence of substances in samples in simulations. If checked, substance residue concentrations are sampled using the correlations between values on the same sample. If unchecked, any correlation between substances is ignored, substance residue concentrations are sampled ignoring the correlations between values on the same sample.
Impute missing values from available values (if unchecked, missing values are imputed with 0)

Boolean

If checked, in procedure of EFSA Guidance 2012, Appendix 1, impute missing values using substance based concentration models. If unchecked, missing values are imputed by 0.
Correlate imputed values with sample potency

Boolean

If checked, in procedure of EFSA Guidance 2012, Appendix 1, correlate high imputed values with high cumulative potency samples. If unchecked, random imputation.
Use occurrence frequencies for imputation

Boolean

Use of occurrence frequencies (e.g., agricultural use frequencies) is relevant for imputation of censored values in the concentration data. Part of the observed censored values and missing values may be imputed with zero when the occurrence frequency is smaller than 100%. If checked, occurrence frequencies are expected as input of this action, otherwise 100% potential presence is assumed for all substances on all foods.

Uncertainty settings

Table 40 Uncertainty settings for module Concentration models.
Name Type Description
Parametric uncertainty

Boolean

For resample concentrations: specifies whether the uncertainty assessment is based on a parametric approach.