Nested case prediction query question

I have a question about what is possible with a prediction query
against a nested table. Say I have a basic customer-product case and nested table mining model like so:

Mining Model DT_CustProd
(
[Id] ,
[Gender] ,
[Age]
[Products] Predict
(
[ProductName] ,
[Quantity]
)
)
Using Microsoft_Decision_Trees

I can write a query to find the probability of product (and quantity) A like so:

SELECT (select * from Predict(Products,INCLUDE_STATISTICS)
where ProductName = 'A' )

FROM DT_CustProd

NATURAL PREDICTION JOIN

(SELECT 'M' AS [Gender],
27 AS [AGE] ) AS t

What if I know that the query customer (M,27) in question has purchased product B, how can I use that in the prediction join to predict product A? The fact that product B was purchased might influence the prediction, right?

Yes, the fact that B was purchased will likely influence the prediction and the model (by marking the table as Predict) actually uses existing products information in predicting new products. The changed query should look like below (a generalized example, for a customer that bought B and C):

SELECT (select * from Predict(Products,INCLUDE_STATISTICS)
where ProductName = 'A' )

FROM DT_CustProd

NATURAL PREDICTION JOIN

(SELECT 'M' AS [Gender],
27 AS [AGE],

(SELECT 'B' AS ProductName, 2 AS Quantity UNION

SELECT 'C' AS ProductName, 1 AS Quantity

) AS Products

) AS t