Filter my DMX Association Algorithm Query

• Tuesday, January 29, 2013 2:05 AM

   

   
  

  0

  Hi,

  I've successfully created a mining structure and Association mining model.  My next step is to run a prediction based on other data in our data warehouse.  I've written the following DMX:

  select flattened T.VisitKey, predict(Procedures, $PROBABILITY, INCLUDE_STATISTICS, INCLUDE_NODE_ID)
  from Procedure_Combinations
  prediction join
  SHAPE
  {
  	openquery([BI Production Dev], 
  				'SELECT  distinct VisitKey
  				FROM dm_Procedures_ToCheck
  				order by VisitKey') 
  }
  
  APPEND
  (
  	{
  	openquery([BI Production Dev], 
  				'SELECT  VisitKey, ProcedureName
  				FROM dm_Procedures_ToCheck
  				order by VisitKey')  
  	} relate VisitKey to VisitKey
  )
  as Procs_To_Check as T
  on Procedure_Combinations.VisitKey = T.VisitKey and 
  	Procedure_Combinations.Procedures.Procedure = T.Procs_To_Check.ProcedureName

  This query returns a row for every customer, regardless of whether or not a rule existed based on their current set of items in their "basket".  How can I modify this query to eliminate the customers that do not have any results?  I tried to add "where $probability >= ###" at the end of the query, but I just get an error.

  Any help is appreciated.

  Thanks,
  Chris

  • Reply
  • Quote