Stepwise regression is an optimization aglorithm aiming to improve the quality of the multiple linear regression by excluding noisy variables.
For models with less than 15 variables, the exhaustive search is the recommended method. Alternatively it is possible to perform "backward search" by starting with all and removing one by one or "forward search" by starting with none and adding one by one variable. Both backward and forward search can have selected variables already pre-included (or pre-excluded). Single step mode allows the user to manually include or exclude individual variables from the model after each step.
| Name | Type | Description |
|---|
| A | TVecList | Holds a list of vectors with data of variables |
| AutoInitBitMask | Boolean | Set this property to false, if you want to manually initialize the BitMask. |
| BitMask | TVecInt | Bit vector holding 1 for included and 0 for excluded vars. |
| BlockAssign | Boolean | Block streaming of specific properties when storing only a "template". |
| EditorActive | Boolean | Returns True, if the component editor is displayed. |
| MaxIter | Int32 | Maximum iteration count before an exception will be raised. |
| Method | TStepwiseMethod | Defines stepwise regression model. |
| OnQualityCriteria | TOnStepwiseQualityCriteria | Implement this event, to return custom quality criteria other than Std. Error to guide the optimization. |
| Reference | TReferenceList | Stores a list of components that have to be notified, when this component is destroyed. |
| reportCoeff | TMtx | Result matrix with size of (IterCount*VarCount) x 5. |
| reportSSE | TMtx | Result matrix size of IterCount x (VarCount + 7). |