Quintiq file version 2.0
|
#parent: #root
|
Method GetSeverityScaleForLowerThresholdViolation () as owning LibOpt_StatisticSeverityScale
|
{
|
Description: 'Return the `LibOpt_StatisticSeverityScale` which holds `LibOpt_StatisticSeverityScaleEntrys` of (`LibOpt_StatisticSeverityScaleEntry.Threshold`, `LibOpt_StatisticSeverityScaleEntry.Severity`) values, which determines what `Severity` to set for a `LibOpt_Issue` that violates the `LowerThreshold` of this `LibOpt_Statistic`.'
|
TextBody:
|
[*
|
// To be overridden in subclasses which:
|
// - `LibOpt_Issues` are created based on the `LowerThreshold` of this statistic, and
|
// - `Severity` for `LibOpt_Issues` are to be derived from a certain scale, rather than using a fixed value.
|
|
/*
|
If the criteria above is applicable for your statistic, then refer to the example below as to how to construct your *severity scale*.
|
|
EXAMPLE OF *LOWER THRESHOLD* VIOLATION SCALE [Ref. 1]
|
|
----------------------------
|
| THRESHOLD | SEVERITY |
|
----------------------------
|
(i) | 15 | 1 |
|
----------------------------
|
(ii) | 5 | 3 |
|
----------------------------
|
(iii) | 1 | 5 |
|
----------------------------
|
|
- The *lower* the `Threshold` violated by the value of a `LibOpt_Issue` of this statistic, the *higher* the `Severity` to set for the issue.
|
Using the the scale in [Ref. 1], the expected `Severity` for some examples are shown below:
|
- Value of issue = 15 --> Severity = 1
|
- Value of issue = 10 --> Severity = 2 // 1 + ( 3 - 1 ) * ( ( 10 - 15 ) / ( 5 - 15 ) )
|
- Value of issue = 1 --> Severity = 5
|
- Value of issue = 0 --> Severity = 5
|
|
- The logic that uses the "severity scale" defined here relies on the following ASSUMPTIONS:
|
(** If any of the ASSUMPTIONS is violated, then the scale might not work as expected. **)
|
- The `Severity` values should be within the inclusive range of [ `LibOpt_Issue::Severity_1_Min()`, `LibOpt_Issue::Severity_5_Max()` ].
|
- No duplicate entries of `Threshold` or `Severities`.
|
- The trend of `Threshold` of entries should be *descending* while the trend of `Severity` of entries should be *ascending*.
|
|
-----------------------------------------------------------------------------------------------------------------------------------------------------
|
|
CODE EXAMPLE:
|
|
>>>>
|
severityscale := construct( LibOpt_StatisticSeverityScale );
|
|
severityscale.Entry( relnew,
|
Threshold := this.LowerThreshold(),
|
Severity := LibOpt_Issue::Severity_1_Min()
|
);
|
|
severityscale.Entry( relnew,
|
Threshold := <Lower Threshold Corresponding With Max Severity>,
|
Severity := LibOpt_Issue::Severity_5_Max()
|
);
|
|
return &severityscale;
|
<<<<
|
|
NOTE:
|
For the <Lower Threshold Corresponding With Max Severity> part, check out the
|
`LibOpt_StatisticSummary::GetLowerThresholdCorrespondingWithMaxSeverityBasedOnOutlierThreshold` helper method to see if it suits your statistic.
|
If so, you can use `this.Summary().GetLowerThresholdCorrespondingWithMaxSeverityBasedOnOutlierThreshold()` for that part.
|
*/
|
|
return null( LibOpt_StatisticSeverityScale, owning );
|
*]
|
}
|
