Dynamic prediction calculations

Question

Dear All,

I have regression model that set in Analysis Service. My model has average inventory,average price and sales unit for t-shirt sales. Sales unit depends on average inventory and average price. I need to predict sales unit based on average inventory and average price.  But I need dynamic calculations. I mean we set up a screen for sales as a windows application C#  and in the backround  the prediction has to be calculated based on given any inventory level and average price. How can I do this? How can I connect Analysis Service and C# screen for dynamic calculations?

Do I need a cube in Analysis service?

Thank you

Answer 1

Have you created data mining model or not yet?

You don't have to create a cube if you already have all the input information for the model (average inventory, price etc).

After the model is created, you can either send DMX query to get dynamic predictions (see sample below) or, if you are using lenear regression, you could just take the formula and put it in the code.

Here is how singleton query should look like

SELECT

    Predict([Model Name].[Sales Unit]) AS [Output 1]      

FROM

    [Model Name]

NATURAL PREDICTION JOIN

(

SELECT

    '29.85' AS [Average Price] , 

    '140' AS [Average Inventory] , 

....

)  AS T

---

Tatyana Yakushev [PredixionSoftware.com]