final_moments

Output moments of the specified type and include them within the set of final statistics.

Specification

  • Alias: None

  • Arguments: None

  • Default: standard

Child Keywords:

Required/Optional

Description of Group

Dakota Keyword

Dakota Keyword Description

Required (Choose One)

Final moment type

none

Omit moments from the set of final statistics.

standard

Output standardized moments and include them within the set of final statistics.

central

Output central moments and include them within the set of final statistics.

Description

When performing a nested study that may employ moment statistics on the inner loop, it can be desirable to control the type of these moments. The final_moments specification supports options of none, standard (default), or central, corresponding to omission of moments, standardized moments (mean, standard deviation, skewness, and excess kurtosis), or central moments (mean, variance, 3rd central, and 4th central).

The presence or omission of moment results in the final statistics influences the outer level mappings in the case of a nested study. For example, final_moments none can allow for a more compact specification of primary and/or secondary response mappings.

Examples

The following method specification overrides the default to print central moments and include them in the set of final statistics.

method,
  sampling
    samples = 50 seed = 1234
    response_levels = 3.6e+11 1.2e+05 3.5e+05
    final_moments central

Overriding the default to none as follows:

method,
  sampling
    samples = 50 seed = 1234
    response_levels = 3.6e+11 1.2e+05 3.5e+05
    final_moments none

allows associated nested model mappings to be simplified from:

model
  nested
    sub_method_pointer = 'UQ'
    primary_response_mapping   = 0. 0. 1. 0. 0. 1. 0. 0. 0.
    secondary_response_mapping = 0. 0. 0. 0. 0. 0. 0. 0. 1.

to a more compact version focused only on the response level mappings (two leading zeros per response function for moment mappings have been removed):

model
  nested
    sub_method_pointer = 'UQ'
    primary_response_mapping   = 1. 1. 0.
    secondary_response_mapping = 0. 0. 1.