.. _method-import_points-tolerance_intervals: """"""""""""""""""" tolerance_intervals """"""""""""""""""" Computes the double sided tolerance interval equivalent normal distribuion. .. toctree:: :hidden: :maxdepth: 1 method-import_points-tolerance_intervals-coverage method-import_points-tolerance_intervals-confidence_level **Specification** - *Alias:* None - *Arguments:* None - *JSON argument key:* None **Child Keywords:** +-------------------------+------------------------+--------------------+--------------------------------+------------------------------------------------------------+ | Required/Optional | Description of Group | JSON Group Key | Dakota Keyword | Dakota Keyword Description | +=========================+========================+====================+================================+============================================================+ | Optional | | | `coverage`__ | The coverage to be used for the calculation of the lower | | | | | | and upper ends of the interval covering the user supplied | | | | | | samples. | +-------------------------+------------------------+--------------------+--------------------------------+------------------------------------------------------------+ | Optional | | | `confidence_level`__ | The confidence level to be used to determine the standard | | | | | | deviation of the equivalent normal distribution. | +-------------------------+------------------------+--------------------+--------------------------------+------------------------------------------------------------+ .. __: method-import_points-tolerance_intervals-coverage.html __ method-import_points-tolerance_intervals-confidence_level.html **Description** The "tolerance_intervals" keyword is used to compute (i) the equivalent normal distribution, given response samples and a desired conficence level between 0 and 1, and (ii) the lower and upper ends of an inverval covering the samples, given a desired coverage between 0 and 1 :cite:p:`Jekel20`. *Default Behavior* By default, coverage is set to 0.95 (95%) and confidence level is set to 0.90 (90%). *Expected Output* Dakota first analyzes the N response samples supplied by the user, and eliminates any sample with a NAN value, resulting in a final amount M <= N of samples to be actually used in the calculations. If M >= 2, then Dakota computes 6 values: the sample mean 'mu', the sample standard deviation 's', the multiplicative factor 'f', the tolerance interval lower end 'mu-f*s', the tolerance interval upper end 'mu+f*s', and the standard deviation 's_equiv' of the tolerance interval equivalent normal. **Examples** .. tab:: Freeform .. code-block:: dakota method sampling samples 5 refinement_samples 4 3 3 sample_type random tolerance_intervals coverage 0.96 confidence_level 0.92 seed = 31415 .. tab:: JSON .. code-block:: json { "method": [ { "sampling": { "samples": 5, "refinement_samples": [4, 3, 3], "sample_type": {"random": true}, "tolerance_intervals": { "coverage": 0.96, "confidence_level": 0.92 }, "seed": 31415 } } ] }