If I have a classification problem where 10 % of my customers churn and 90 do not churn, should I build traing sets for my models that reflect this reality? Or is it best to give my model a training set with 50 % of each cases. Will this give my model enough information so that ...

I've build program that needs to update data, in very complex way, so i had
to use cursor for use it.
The cursor works fine. But i need also to update the current record of the
cursor

Errors in the back-end database access module. OLE DB was unable to convert a value to the data type requested for column 9. Errors in the OLAP storage engine: An error occurred while processing the 'Suppliers Fact' partition of the 'Suppliers Fact' measure group for the ...

I have documentation Program as part of client data.
In the program each user can store files for each client. The files are
mostly TIFF, and in some times PDF, or XLS, DOC.
Today i store the files in the server.

I have a problem with convincing my coworker on using foreign keys in
database.
He doesn't want to use foreign keys. His arguments are: They're not
nesessary as long as you

I am just starting to use the SQL Server 2000 Analysis Services. I have
created a data source but i did not save the password as i was "told" by the
system that "it's unsafe as the password stored will be unencrypted"
However, when i tried to create a cube , it says login failed ...

I'm running the time series algorithm on the "Adventure-Works" Cube (specifically on dimension "Product" key "Subcategory", time dimension "Date" key "Time Key", measure "Internet Sales Amount").
I'm getting in many cases standard deviation "1".
This is a suspicious value.
Does any ...

Is it possible in someway to execute some DMX statement against a trained model and recive a "score" or a measure, or anything, that describes how well the model will perform predicting?
I would like to be able to compare many models using code and DMX,
/Erik

I am creating mining structures and mining models using DMX language.
First I create Mining model, for example:
CREATE MINING MODEL [myModel]
( list of attributes, content types, data types...) USING SomeAlgorithm (AlgParameters)

Is it only possible to execute SELECT DMX statements from SQL via 'OPENQUERY'? I would like to do things like this;
SELECT * FROM OPENQUERY(ASLOCAL,
'DROP Mining Structure sGLDetCostDebet')
But I get this error message;

Are there any examples, information to build a SSIS package to process the cubes/mining models ?

Platform: SQL Server 2005 Analysis Services on W2K3
While creating regular databases, I know how to change the storage location of database files.
By default SSAS2005 points
"C:\Program Files\Microsoft SQL Server\MSSQL.2\OLAP\Data"

Every time I try to do anything even to the tutorial db's, I get an error
saying that I cannot save because the database or object is read-only. Also,
I've added my own db and a datasource, but it does not save my dimensions or
anything, and I cannot connect via the datasource ...

Got the following error while creating Time Dimension.
Using W2K3, SQL Server 2005, Visual Studio 2005.
Source database and target SSAS Database are all on my local workstation.
Unlike treditional Visual Studio projects, I do not see add references option for Analysis Services ...