Re: Scripting data mining

From: Bogdan Crivat [MSFT] (bogdanc_at_online.microsoft.com)
Date: 02/12/05

  • Next message: Stefan Rosi: "Re: Paging in SQL Server, an overview of some available approaches" 
    Date: Fri, 11 Feb 2005 20:19:29 -0800

    SqlConnection only works with SQL Server.
    For Analysis Services you should use the generic OleDb classes in ADO.Net:
    System.Data.OleDb.OleDbConnection and System.Data.OleDb.OleDbCommand
    The connection string should look like:
    "Provider=MSOLAP; Data Source=<Your Server Name>; Initial Catalog=<Your
    Database>";

    A query analyzer-like tool for Analysis Services Data Mining is available as
    a sample at:
    http://www.sqlserverdatamining.com/DMCommunity/SQLServer2000/Links_LinkRedirector.aspx?id=109

    If you have Analysis Services 2000 installed, most likely you have the
    FoodMart 2000 sample database installed as well.
    Once you connect to that database, you can issue a query like below, for
    test purposes:

    select [member card] FROM [Member Card RDBMS] NATURAL PREDICTION JOIN
    (SELECT 'M' AS [marital status]) AS T 

    -- 
--
--
This posting is provided "AS IS" with no warranties, and confers no rights.
Please do not send email directly to this alias. It is for newsgroup
purposes only.
thanks,
bogdan
"B.J." <BJ@discussions.microsoft.com> wrote in message
news:DE7ADB20-4549-4D14-AA55-78C30F2B4824@microsoft.com...
> Thank you. I have one additional question (becuase I never worked with MS
SQL
> Analysis server, only with MS SQL 2000 Database server).
> I want to code it in C#.NET. Should I used classical SqlConnection and
> SqlCommand objects (as when connecting relational database) also for
> connecting to analysis server ?
> And can I somehow run this query in Query Analyzer or some other tool
> respectively in Visual Studio .NET for quick testing ?
>
> Yet one more thanks.
>
> "Bogdan Crivat [MSFT]" wrote:
>
> > I assume you are using Data Mining coming with Microsoft Analysis
Services.
> > Yes, you can do that. Open an OLEDB (or ADO, or ADO.Net) connection to
the
> > Analysis Services server.
> > On that connection, you can execute INSERT INTO to fully reprocess a
data
> > mining model
> > A singleton prediction query is most likely the best solution for
predicting
> > based on client input:
> >
> > Something like:
> >
> > SELECT [College Plans] -- the column to predict
> > FROM [CPModel] -- the model to predict from
> > NATURAL PREDICTION JOIN
> > ( SELECT 50000 AS [Parent Income], -- client input here
> >                  120 AS [IQ] -- client input here
> > ) AS T
> >
> > -- 
> > --
> > --
> > This posting is provided "AS IS" with no warranties, and confers no
rights.
> > Please do not send email directly to this alias. It is for newsgroup
> > purposes only.
> >
> > thanks,
> > bogdan
> >
> > "B.J." <BJ@discussions.microsoft.com> wrote in message
> > news:9E901559-13E5-4B72-BFA6-6F3CA98E92DA@microsoft.com...
> > > Hi,
> > >
> > > Can I execute from code refreshing of data and full processing of
Mining
> > > Model ?
> > > And also execute prediction query on this mining model with client
input ?
> > >
> > > Thank you
> >
> >
> >
  • Next message: Stefan Rosi: "Re: Paging in SQL Server, an overview of some available approaches"

    Relevant Pages
    Loading