Note
Access to this page requires authorization. You can try signing in or changing directories.
Access to this page requires authorization. You can try changing directories.
Predicts associative membership.
Note
This is different from creating a prediction on an association model.
Syntax
PredictAssociation(<table column reference>, option1, option2, n ...)
Applies To
Classification algorithms and clustering algorithms that contain predictable nested tables. Classification algorithms include the Microsoft Decision Trees, Microsoft Naive Bayes, and Microsoft Neural Network algorithms.
Return Type
<table expression>
Remarks
The options for the PredictAssociation function include EXCLUDE_NULL, INCLUDE_NULL, INCLUSIVE, EXCLUSIVE (default), INPUT_ONLY, INCLUDE_STATISTICS, and INCLUDE_NODE_ID.
Note
INCLUSIVE, EXCLUSIVE, INPUT_ONLY, and INCLUDE_STATISTICS apply only for a table column reference, and EXCLUDE_NULL and INCLUDE_NULL apply only for a scalar column reference.
INCLUDE_STATISTICS only returns $Probability and $AdjustedProbability.
If the numeric parameter n is specified, the PredictAssociation function returns the top n most likely values based on the probability:
PredictAssociation(colref, [$AdjustedProbability], n)
If you include $AdjustedProbability, the statement returns the top n values based on the $AdjustedProbability.
Examples
The following example uses the PredictAssociation function to return the four products in the Adventure Works database that are most likely to be sold together.
SELECT
PredictAssociation([Association].[v Assoc Seq Line Items],4)
From
[Association]
See Also
Reference
Data Mining Extensions (DMX) Function Reference
Functions (DMX)
Mapping Functions to Query Types (DMX)